Handbook of Psychology, Volume 4: Experimental Psychology

74 Mood, Cognition, and Memory

To elaborate, consider first the conventional list-learning
paradigm, alluded to earlier. According to Bower and Forgas
(2000, p. 97), this paradigm is ill suited to demonstrating
MDM, because it

arranges only contiguity, not causal belonging, between presen-

tation of the to-be-learned material and emotional arousal. Typi-

cally, the mood is induced minutes before presentation of the

learning material, and the mood serves only as a prevailing back-

ground; hence, the temporal relations are not synchronized to

persuade subjects to attribute their emotional feelings to the ma-

terial they are studying. Thus, contiguity without causal belong-

ing produces only weak associations at best.

In contrast, the model allows for strong mood-dependent
effects to emerge in studies of autobiographical memory,
such as those reported by Eich et al. (1994). Referring to
Figure 3.2, which shows a fragment of a hypothetical asso-
ciative structure surrounding the conceptlake,Bower and
Forgas (2000, pp. 97–98) propose the following:

Suppose [that a] subject has been induced to feel happy and is

asked to recall an incident from her life suggested by the target

wordlake. This concept has many associations including several

autobiographic memories, a happy one describing a pleasantly

thrilling water-skiing episode, and a sad one recounting an

episode of a friend drowning in a lake. These event-memories

are connected to the emotions the events caused. When feeling

happy and presented with the list cue lake,the subject is likely

(by summation of activation) to come up with the water-skiing

memory. The subject will then also associate the list context to

the water-skiing memory and to the word lakethat evoked it.

These newly formed list associations [depicted by dashed lines

in Figure 3.2] are formed by virtue of the subject attributing

causal belonging of the word-and-memory to the experimenter’s

presentation of the item within the list.

These contextual associations are called upon later when the

subject is asked to free recall the prompting words (or the mem-

ories prompted by them) when induced into the same mood or a

different one. If the subject is happy at the time of recall testing,

the water-skiing memory would be advantaged because it would

receive the summation of activation from the happy-mood node

and the list context, thus raising it above a recall level. On the

other hand, if the subject’s mood at recall were shifted to sad-

ness, that node has no connection to the water-skiing memory

that was aroused during list input, so her recall of lakein this

case would rely exclusively upon the association to lake of the

overloaded, list-context node [in Figure 3.2].

Thus, the revised network model, like the AIM, makes a

clear case for choosing autobiographical event generation

over list learning as a means of demonstrating MDM.

Nature of the Retrieval Task

Moreover, both the AIM and the revised network model ac-

commodate an important qualification, which is that mood

dependence is more apt to occur when retention is tested in

the absence than in the presence of specific, observable

LIST

CONTEXT

WATER

SKIING

MEMORY

HAPPY

EMOTION

SADNESS

LAKE EMOTION

FRIEND’S

DROWNING

MEMORY

Figure 3.2 Fragment of a hypothetical person’s associations involving the concept of lake. Lines represent associa-

tions connecting emotion nodes to descriptions of two different events, one happy and one sad. The experimental con-

text becomes associated to experiences that were aroused by cues in that setting. Source:Bower and Forgas, 2000.