[page 18, see Gleitman, 1999; pp. 148-150 or Gleitman, 2004, p. 150-2: or Lieberman, 2000; pp. 115-120]
Garcia and Koelling (1966)

Ss. 10 rats in each of 4 groups, individually tested in a small box with a drinking spout, in which every lick of the spout could be counted.


1 week of habituation to drinking plain water in the box and pre-tests “bright-noisy” water (flashing light and clicks when the spout was licked) and “tasty” water (with flavours added).

Then Training: e.g with lithium. One 20-min session every 3 days in which rats could drink only water with added lithium chloride (which they cannot distinguish from salty water), with the “bright-noisy” stimuli also present. These were interpolated with days when the rats were tested with plain water – no extra stimuli and no aversive effects.

A group was run for which the floor was briefly electrified 2 sec after licks at the spout and a group where similar shocks were delayed for longer, to produce a decline in drinking of the same kinds as that seen in the lithium group

Testing. After 2 days rest, without any further aversive effects, subgroups of Ss were tested either with water accompanied by just the audio-visual cue, or water with just the flavour cue.


Ss trained with gastric distress (either due to lithium, or X-irradiation) subsequently suppressed drinking of the flavoured water much more than the water accompanied by audio-visual cues used in training.

Conversely the rats which had received peripheral aversive stimulation in training suppressed drinking in the presence of the audio-visual cue, but not with the flavour cue.

Gleitman et al table

Back to - Phenomena showing that stimulus-pairing (contiguity) is not the only factor in Pavlovian conditioning. [bottom of page 17]