Basic level, Superordinate and Subordinate.

For example:

Basic- Bird

Superordinate- Animal (More general than basic)

Subordinate- Rosella (Most specific level, in this example it is the bird's name).

Concepts= General ideas about the classes of objects

Categories= Classes for objects represented by concepts

Categorisation= Act of saying that some thing belongs to a larger category of things

Recognition= Act of saying that some thing is exactly the same as something you have previously encountered

1. Categories reduce cognitive complexity- reduces the number of things we have to remember

(Example: Email can be read more efficiently if sorted into categories)

2. Categories allows us to generalise in new situation

(If you have a category to guide your actions, you can zero in on what you are looking for and ignore irrelevant details)

Equivalence class-

Distinguishable stimuli treated as same thing once placed in same category.

Useful when we need to make connections between objects that have different apparent form

Cues and features of stimuli guide us to make categorization judgments, and how we categorize stimuli determines how we behave towards or with those stimuli. 

Unidimensionalrule

•Big vs Small

•Only one dimension necessary, very easy to learn

Conjunctive Rule (AND rule)

•Example: Big & Square

•2 features combined =Two dimensions necessary= Need 2 features in order to group an item into a category

Classical categorisation relies on rules like disjunctive, unidimensional or conjunctive rules.

It makes it difficult to define what the necessary conditions in order to categorise things in the same category.

For eg. Games category can refer to an activity that requires players and a goal. This means that chess and basketball are both games even though they are both very different activities.

Hence, problem with classical categorisation is that anything that meets the requirements of the rules automatically becomes part of the same category.

Typicality

If you use the same strict rules an apply it to every item you categorise, you may not be able to categorised items that are not typical of the items in a category.

For example, if fish was defined as a streamlined torpedo shaped creature, eels will not fit into that typical rule and will not be categorised as a fish even though it cleary is a fish.

The typicality of an item with the rest of the category members is not the criteria on which one places an item in that category.

The more typical an item was however, the higher number of attributes that it shared with other objects.  

The highly typical items contained the features shared by lots of different category member. There were also many objects made up of idiosyncratic features that were not typical of the category. These atypical items however were also category members.

If we use similarity when categorising items, then some items will be typical and some items will not.

Items which are highly dissimilar have a high psychological distance

once two things are far enough apart they are treated as completely different objects (e.g., no generalization).

How do we use protoypes in categorisation?

What can we infer from Posner's study with dot patterns?

we store some average of all of the category members that we’ve seen before as a prototype. We then compare all the features of the new items to the prototype

Posner's study:

• Subjects were presented with dot patterns which were all highly distorted versions of the prototype dot pattern.
  
• To create the high distortions, Posner took the prototype and moved the dots around a lot so that the high distortions weren’t really the same as the prototype but they had an underlying family resemblance.
  
• Subjects viewed 5 of these highly distorted patterns and were told that they were all members of the same category.

• the prototype was not shown.

during the test phase,subjects were shown

• the high distortions that they had seen before,
  
• new high distortions,
  
• new “low distortions” which were created in the same way as the new high distortions but the dots weren’t jiggled around as much.
  
• And random patterns which weren’t created from the prototype at all.

They had to respond “Yes” if they thought the pattern was a member of the category that they just saw, or “No” if it was not a member of the category that they just saw.

Result:

• people respond yes 90% of the time when they are shown one of the old high distortions that they saw during training.

• They also say yes most of the time when they see the prototype, even though the prototype was never shown during the training phase.

• The proportion of Yes responses drops off for new low and high distortions. (Participants were still able to categorise these items as belonging to the category, even though they were significantly lower than old high and the protoype)
  

• And finally decreases to around 10% for random dot patterns.

Inference:

even though the prototype and other low distortions weren’t shown during the training phase, people know that this items are members of the category represented by the old high distortions.

What is the criticism of prototype? What is the alternative explain how we categorise items?

Abstract prototypes are not psychological plausible.

Alternative explanation- Exemplar model:

We store all of the members of the category along with their specific category label.

When we test items for similarity, we test the items against everything we have seen before.

When categorising we may focus on what specific feature (selective attention) and use that as a basis to assess whether new items proposed to us belong to the category.

This happens for example, when we are given 2 new items and asked to determine which items is more similar to the category members.

(See the Dalmatian and chihuahua example, slide 77)

Recognition is the ability to explicitly determine whether an item is identical to a previously experienced exemplar or event.

Recognition is determining whether a presented item is identical to the other exemplars seen in the past. (Do I remember seeing this item before?)

Recognition has a higher and stricter criterion than categorisation.

Must have seen item before item meets the recognition criteria.

VS

Categorisation is determing whether a presented item is similar enough to belong to the same category as the exemplars seen in the past.

(Does this item belong to the same category?)

Lower criterion= Does it belong to the same group? (Item does not have to be similar enough for you to put it in the same category).

Knowlton and Squire (1991)-

Declarative system in amnesics is damaged, their old/new recognition is impaired. However, the implicit system is intact so their categorization performance does not suffer.

Therefore an inference that separate parts of the memory system

and different neural substrates were used for recognition and categorisation.

Dual Systems theory:

Recognition uses- Declarative/ Explicit Episodic Memory

Categorisation uses- Non-declarative/ Implicit Procedural Memory

The difference in performance in recognition and categorisation may reflect the application of different criterion for recognition and categorisation task.  

Hence, it does not mean a difference in performance levels means that different memory systems are used for categorisation and recognition

Counter proposal:

Single-system theory:

How does the theory of theory or prior knowledge affect the categorisation process?

Hardiman, Dufresne & Mestre (1989)