Francis Galton, the English eugenicist who wro...Galton - (1880) - had people imagine their breakfast tray from that morning. He noted some fascinating characteristics of their descriptions.

(1) Illumination - dim or fairly clear? Brightness comparable to actual scene? People report their mental image is dimmer than if you were really there.

(2) Definition - everything well defined at the same time or does the sharpness vary. Most report it varied

(3) Color - are the colors distinct and natural or were they washed out?

McKellar (1972): 97% of adults report visual mental imagery. 3% couldn’t do it. Not everyone seems to have visual mental imagery.

Imagery and Memory

Dual coding hypothesis (Paivio)

Concrete concepts (chair, rock) are stored in memory twice: once in semantic code (meaning), one in analog code (mental image). You have two different routes of arriving at destination. Twice as likely to do so. Two fish in the pond.

Abstract concepts (justice, cognition) are stored in memory just once, in semantic code because it’s difficult to form a mental image of them. You can do it metaphorically by making a statue (supreme court statue of blind woman holding scale and sword).Explains why concrete concepts are better recalled than abstract concepts in list learning experiments.

The imagery controversy

(1) analog position: mental images are very similar to visual images. Mental images are pictures in the head. Mental and visual imagery make use of the same cognitive processes.

Three lines of evidence:

a. mental rotation (shepard and Metzler) - subject shown pairs of 3D objects: one is rotated in relation to the other. The subjects were asked to judge if the two objects are the same as fast as possible. Dependent Variable = reaction time. Independent variable is the angle of rotation. Graphed data was a perfect line. between 0 and 20 degrees took 1 to 2 seconds. 180 degrees took about 5 seconds. The more the degrees an object is turned the longer it took to identify.

S&M: Functional equivalence between objects we see and objects we imagine. We manipulate things in the exact same way.

Problem is mental rotation is not truly a memory task.

b. Mental Scanning (Kosslyn, Ball, & Reiser) - the Island Study, subjects were asked to memorize a map of an island with a number of marked locations. Map then taken away and they’re asked to shift their attention from one location to another. Mental scanning data graphically is linear as well. The further away two points on the map were the longer it took the subject to respond.

c. Mental Comparisons (Moyer 1973) - Which is bigger, a mouse or a cat? What she did in experiment was vary the difference in size between the two objects she would ask about. Which is bigger a cat or a dog? As the estimated difference in animal size goes up, the reaction time goes down.

(2) Propositional position - information stored in propositions in LTM. Propositions are pieces of semantic code. Some attribute that an object has.

Mental images are the by products of retrieving propositions from memory.

Advocated by Pylyshyn.

imagine an apple. How did you do it?According to analog position you retrieved it from LTM and got apple.

According to propositional position: we have to build an apple. What makes an apple an apple. Retrieve relevant propositions from LTM (is red) (is shiny) (is round) (has a stem) and we have an apple. The two theories differ crucially in what is being involved.

Arguments for it:

Storage of mental images costly (similar to templates)

In some cases, mental images are different than visual images (house example from earlier)(dimmer, less distinct, less colorful)(apple example – was it a real particular apple or an idealized apple)


Image by Fuuu via Flickr

Cognitive Maps

Internal representations of the environment.

ask students to draw maps of their college campus. Most can consistently place landmarks on campus. These are tall or large buildings that can be seen from a distance. They consistently place paths that are frequently used.

People often makes mistakes though

(1) size of familiar locations is exaggerated (compared to unfamiliar locations)

(2) Curved paths are drawn as if they were straight (Central Avenue)

(3) distance between familiar locations is underestimated (home to campus) and distance between unfamiliar locations is over-estimated (Memphis and an unfamiliar city)

How many miles from Memphis to New Orleans?

How many miles is it from Memphis to Minneapolis?

Geography questions:

Which city is further west: Reno, NV or Los Angeles, CA? Reno is further west

Which city is further east: Miami, Fl or Lima, Peru? Hint: Peru is on the west coast of South America. Lima, Peru is actually further east.

Analog position: Maps in the head should be relatively accurate. Can’t easily predict the errors we’ve seen.

Propositional position: errors should reflect propositions in LTM.

Reno (in Nevada)

CA (west of Nevada)

LA (in CA)

LA (on coast of CA)

Problem is that California is also south of Nevada.

Related articles