NAB-Categorisation

Concepts are not always defined by specific features, sometimes concepts do not have necessary or sufficient features to define them.

Types of Concept:

1. Basic level concept--based on similarity of perceptual qualities

2. Superordinate concept--groups of basic level concepts; not based on perceptual similarity

3. Abstract concept--does not refer to individual entity, but to some property, relation or state

SO... Can animals form basic level concepts? superordinate concepts? abstract concepts? If yes, how they do it? Do animals form concepts in the same way as humans?

=>Bhatt, Wasserman, Reynolds & Knauss (1988)-pigeons were trained to peck different keys for exemplars of each category of item presented. When new exemplars are tested with, they were able to respond correctly with an accuracy of 60% than the 80% of trial items. This suggests that the pigeons had formed a concept of the tested items.

Theories of basic level concept formation:

i) Exemplar theory: Learn about every instance independently. Classify novel exemplars on the basis of similarity to learned instances

--Animals are more accurate with the novel test stimuli as they are storing information about the training exemplars.

ii) Prototype theory: Abstract a prototype that corresponds to the central tendency of training exemplars. Classify novel exemplars on basis of similarity to new prototype.

--Human categorize the prototype more accurately than th etraining stimuli, even if it has never been encountered before.

SO do animals and humans store concepts in different ways? 

=>Aydin & Pearce (1994)-trying to show a prototype effect in animals. The birds learned that three positive patterns were always paired with food, where the three negative patterns were not. Birds pecked more at positive than negative patterns. When one of the three elements in a set is subsituted by an element from the other set, the birds responded more to the typical +ve patterns than the substituted ones, and less to the typical -ve patterns than the substitued ones. It was an evidence of protitype effect.

SO do humans store examplars as well?

=>Whittlesea (1987) created 3 lists of 5 words accordig to a prototype where all 3 lists differ from the prototype by 2 letters, but list 1 is more similar to list 2 than list 3. Paticipants studied list 1 and were tested with all 3 lists. If they have abstracted the prototype, then they should be uqually good at categorising the lists (randomly) as they all differ from the prototype by 2 letters.

But if they are remembering exeplars, then list 1 should be the easiest (studied), then list 2 (differs a little from 1) and 3 (differs a lot from 1).

**Humans show results consistent with th exemplar theory.

Both humans and animals retain information about the training items/exemplars. The prototype theory turns out to be explainable by a variation of examplar theory!

If the exemplar theory assumes that each stimulus comprises a set of component feature that are associated with category membership, then it can explain the prototype effect.

The exlanation is viewed as a new theory: "feature theory".

Feature theory and exemplar theory both say that you store something about the stimuli on each trial. 

Exemplar story says that new stimuli are classified based on the similarity to stored examplar; while the feature theory are classified based on sharing features that belong to category.

Some argues that categories are formed by means of associative learning where features of category become associated with thee category label. Blocking is a key characteristic of associative learning

=>Shanks (1990)- subjects were given trials where medical symptoms paired with disease diagnosis. They were provided with two diseases: one common while another rare, and three symptoms including one targeted symptom. Blocking is formed with lesser trials between rare disease and the targeted symptom.

Association between stimuli only forms if the category is surprising. A rare disease is more surprising in this case.

This suggests that associative learning is the best explanation of categorization task in humans.

Can animals form superordinate categories?

=>Wasserman, De Voider & Coppage (1992)- trained pigeons to make response A to category 1 and 2, and to make response B to category 3 and 4. Then, they were trained to make response C to category1 and response D to category 3. They were later tested with category 2 and 4, with choice of response C and D. They made response C to category 2 and response D to category 4.

They seem to treat category 1 and 2 as equivalent because of the earlier paring with response A in training phase,

Is this thw same as how human do it?

It is argued that this is not true categorization that seen in humans, but just simple associative learning.

Abstract concept formation in animals?

=>Wasserman, Hugart & Kirkpatrick-Steger (1995)- showed pigeons complex stimulus displays and given a choice of red or green key. The pigeons were rewarded for pecking the red key when a picture of same items are schown and green key for different items shown in a picture. They seems to be able to peck on the correct key when different arrays of a different specific icons.