It's an ordinary afternoon at Copenhagen Central Station. At 2:32pm, a man who appears to be a run-of-the-mill street performer sets up a drum in the center of a large hall. A cellist joins him. A woman approaches with her flute. The melody is sort of recognizable... It sounds sort of like Ravel's Bolero. Pretty cool jam session, right?

Then the clarinet and bassoons and all the rest of the instruments start playing. People pull out their cell phones and record video. Fathers and children take a seat on the tile floor to listen. Mothers with strollers slow down to watch. Within minutes, its like an entire symphony orchestra has assembled itself in the center of the station. Actually, it's exactly like that: it's the Copenhagen Philharmonic! And they are playing Ravel's Bolero!

Once the piece is done, the musicians disperse, and its business as usual at the station. It's as if it never happened.

Why is this musical flashmob so interesting? There's something unique and engaging about this. This is a very different experience than watching an orchestra perform in a music hall. Perhaps it's because of the jarring disconnect between the familiar experience of hearing classical music and the unexpected context within which that experience occurs.

You don't have to go to Copenhagen and wait for an orchestral flashmob to surround you in order to experience this sort of disconnect. You can get it at the liquor store, or indeed, at most college campuses.

Four Loko was a fruit-flavored, caffeinated, alcoholic drink that was invented by three Ohio State University students in 2005. Following a series of accidents, injuries, and deaths on college campuses and elsewhere, most of the discussion about the harmful effects of the drink centered on the combination of caffeine and alcohol. "Some have claimed," writes McMaster University psychologist Shepard Siegel in the latest issue of Perspectives on Psychological Science, "that the stimulant [the caffeine -JGG] masked the intoxicating effects of alcohol, thus encouraging excess alcohol consumption."

The FDA declared that it would be illegal to add caffeine to alcoholic beverages, and on November 18, 2010, sent a letter to several companies including the manufacturers of Four Loko saying that such beverages were unsafe and could lead to "hazardous and life-threatening situations." Within fifteen days, the beverage manufacturers had removed all the caffeine from their alcoholic products.

Case closed? That Four Loko was especially dangerous is clear. But was caffeine the culprit? Siegel doesn't think so.

For one thing, caffeine doesn't seem to affect the way that alcohol gets absorbed by the body. Several studies showed that the concentrations of alcohol in breath and blood are similar following consumption of alcohol alone or together with caffeine. Behavioral studies of caffeine-alcohol interactions are more mixed; some report that caffeine can offset the negative effects of alcohol, while others found no effect of caffeine. In self-report studies, adding caffeine to alcohol has increased, decreased, or had zero effect on subjective drunkenness.

One interesting study had participants drink coffee just after drinking alcohol, and then asked them to complete a motor task. What they found was that performance was worse only among the participants who were told by experimenters that caffeine can counteract the negative effects of alcohol. Participants who were told nothing but still drank the coffee and alcohol had better performance, relative to the other group. Furthermore, the same effects were observed when the experimenters used decaffeinated coffee instead of the regular caffeinated. These sorts of results are due to the ironic effects of expectancy. It is possible that those who did worse on the task didn't try to compensate for the impairing effects of alcohol, expecting the caffeine to do the work for them whether or not they actually had caffeine in their systems.

When Four Loko argued in a public statement that their caffeinated alcoholic beverage was no different from "having coffee after a meal with a couple glasses of wine," they might not actually have been completely wrong. Pharmacologically speaking, Siegel points out, they were probably right. At least with respect to the caffeine-alcohol combination.

However, many drugs (including alcohol) are known to be more potent if they are taken in an unusual context, rather than in the same environment in which they are usually taken. When consuming alcohol in ways that are not typical for alcohol consumption, its effects are intensified. Instead of the usual tolerant response to a drug, where a user needs more of the substance in order to get the equivalent effect, a larger response occurs. In a 1976 paper in Science, Siegel termed this the situational specificity of tolerance.

Environmental variables ranging from the room where a drug is administered to ambient temperatures to magnetic fields may influence an individual's drug-related tolerance. Siegel cites several studies that demonstrate situational specificity particularly when it comes to the lethal effects of drugs. Addicts who have become tolerant to otherwise lethal amounts of a given drug (such as opiates) may experience an overdose if they take their typical dose in an atypical setting. These results have been found in species ranging from rats and mice to humans. Critically for the case of Four Loko, flavor cues can also modulate the specificity of tolerance.

Could it be that Four Loko is so effective at producing drunkenness because it doesn't look or taste like alcohol? In 1990, McCusker and Brown conducted a study of college students, all of whom were "experienced, social drinkers" - on average, they drank about half a pint of beer each day. One group was given alcohol in the traditional setting: a simulated bar. The second group was given the same dose of alcohol in a new environment and context: it was combined with sweetened soda water and administered in an office setting. Behavioral and physiological measures showed that the second group became more intoxicated than the first group.

What this comes down to is classical (Pavlovian) conditioning. If a dog learns that there is an association between meat powder and the sound of a bell, then the sound of the bell alone will become sufficient to induce salivation in the dog. The salivation indicates that the dog is preparing for the arrival of the meat powder. Similarly, a social drinker inevitably learns the association between the effects of alcohol and the environment in which they drink. At the same time, the body learns to prepare for the alcohol, and it begins to do so in response to the environment, before the alcohol is even ingested. The net effect of the alcohol therefore decreases over time, which leads the drinker to drink more.

Taken together, Siegel's argument is convincing: people become especially drunk after drinking Four Loko because of the unexpected way in which it is presented: it doesn't actually taste like alcohol. The caffeine probably isn't the problem at all!

If Siegel is right, the sans-caffeine approach that the manufacturers of Four Loko have taken should be even more troubling. On January 4, 2011, they issued a press release describing their new Four Loko XXX Limited Edition: "a brand new flavor profile ever four months." Once someone becomes tolerant to the effects of the alcohol in the green apple flavor, which began sales in January, their tolerance would be eliminated when the blueberry lemonade flavor was released in May.

While many were wringing their hands over caffeine, New York Times food writer Frank Bruni (twitter) came a lot closer to the real reason for the overinebriation associated with Four Loko. He wrote:

...what I quickly came to see was that if you set out to engineer a booze delivery system that is as cloying, deceptive and divorced from the usual smells, tastes and presentation of alcohol as possible, you’d be hard pressed to come up with something more impressive than Four Loko.

Intentional or not, Four Loko works because it takes advantage of the situational specificity of tolerance, perhaps better termed The Four Loko Effect. It has more in common with the Copenhagen Philharmonic Flashmob than with your morning cup 'o joe.

Siegel, S. (2011). The Four-Loko Effect Perspectives on Psychological Science, 6 (4), 357-362 DOI: 10.1177/1745691611409243

McCusker CG, & Brown K (1990). Alcohol-predictive cues enhance tolerance to and precipitate "craving" for alcohol in social drinkers. Journal of studies on alcohol, 51 (6), 494-9 PMID: 2270057

Siegel S (1976). Morphine analgesic tolerance: its situation specificity supports a Pavlovian conditioning model. Science (New York, N.Y.), 193 (4250), 323-5 PMID: 935870

Four Loko photo via Wikimedia Commons.

For more on Four Loko:

Alcoholic Energy Drinks, by David Kroll

Four Loko for locos, also by David Kroll