In today's experiments, one of the most common tricks that experimental psychologists play is giving people placebos instead of drugs. In the past, it was pretty much the reverse. Doctors would call people in, tell them they'd get an innocuous substance, and then inject them with something a thousand times more powerful.
In 1962, experimental subjects were told they were getting a vitamin shot, because doctors wanted to assess the effect of vitamins on visual acuity. Instead, the subjects were given a shot full of adrenaline. They were then taken into a room where a fellow subject — actually a member of staff who was helping with the experiment — was goofing around. The subject and the undercover staff member were soon climbing around on the furniture and throwing paper airplanes at each other. The adrenaline was making the subject get the shakes, as well as a fast heartbeat and a flushed face, but they reported having a great deal of fun and feeling nearly euphoric.
Other subjects did the same thing, with the same fellow goof-off, but didn't feel the pleasure and happiness. They just felt their body shaking and flushing, and their pulse racing.
The difference was that the non-euphoric people were told that the "vitamins" would cause their physical symptoms. They were told to expect physical reactions like facial flushing and rapid heartbeat. They knew that their body was reacting to physical stimulus. The subjects who were told the vitamins should have no effect on them felt the physical effects of an adrenaline high, in a situation where they would expect to be happy, and assumed that they were so very happy that they felt high.
In other words, when our body feels enough of a physical sensation — without a physical explanation -—our mind attributes that physical sensation to an emotional state. If we are given the physical symptoms of being in love (pulse pounding and head spinning) we assume we're in love. If we are given the physical symptoms of feeling sad (watering eyes and a sick feeling in the stomach) we assume we are sad. We adjust our emotional state to fit our physical symptoms — but only in the absence of a more reasonable explanation.
[Via Cognitive, Social, and Physiological Determinants of Emotional State]