Lecture 2 - Conditions of Learning

Question 1

1. Taken in isolation, the experiments considered in Lecture 1 suggest that we need only arrange that two things happen close together in time for their memories to become associated. That is, temporal contiguity is both a necessary and sufficient condition for associative learning to occur.
2. In this lecture we shall consider evidence indicating that temporal contiguity is not a sufficient condition for learning to take place.
3. We will then examine one influential formal model of learning, the Rescorla-Wagner model, that provides a simple way of accounting for the conditions under which learning does and does not occur.

Answer 1

Understand?

Question 2

Name three effects that have been especially important in shaping our analysis of associative learning.

Answer 2

• overshadowing
• blocking
• contingency

Read blue text on page 1 of G doc notes.

Question 3

Describe this study:
Overshadowing in a conditioned suppression procedure (Mackintosh, 1976)

Answer 3

See blue text on page 1-2 of G doc notes.

Question 4

Describe this study:
Blocking in a conditioned suppression procedure (Kamin, 1968, 1969)

Answer 4

See blue text on page 2 of G doc notes.

Question 5

Describe this study:
Contingency effects in a conditioned suppression procedure (Rescorla, 1968)

Answer 5

See blue text on page 2 of G doc notes.

Question 6

What is the role of surprise in associative learning?

Answer 6

See green text on page 3 of G doc notes.

Question 7

Describe an informal account of overshadowing in terms of the role of “surprise”.

Answer 7

See green text on page 3-4 of G doc notes.

Question 8

Describe an informal account of blocking in terms of the role of “surprise”.

Answer 8

See green text on page 4 of G doc notes.

Question 9

Describe an informal account of the effect of contingency in terms of the role of “surprise”.

Answer 9

See green text on page 4 of G doc notes.

Question 10

Describe the Rescorla-Wagner model (1972).

Answer 10

See purple text on page 4-7 of G doc notes.

Question 11

Conclusions:
1. We began with a series of related effects (overshadowing, blocking, and contingency) that show associative learning is not simply driven by contiguity as had been thought since the early associationists (e.g., Mill, 1843).
2. These failures of contiguity to result in learning can be taken to indicate that the US has to be surprising if it is to produce learning.
3. The model developed by Rescorla and Wagner (1972) captures the informal notion that “surprise” influences associative learning within a precise formal model that, as we shall see next week, makes novel and testable predictions.

Answer 11

Understand?