side effect

When something causes both an effect that was intended and some other effect as well, the unintended effect is a side effect. For example, the intended effect of the antibiotic erythromycin is to kill bacteria, but a side effect for me is that it makes me vomit.

"Side-effect" or "side effect" is a relatively recent phrase. It first appeared in a book title with Lewis Levin's Side Effects of Medication (Die Niebenwirkungen der Arnzeimittel) in 1881. The OED's first citation is from 1884 for the general meaning, and they didn't list the medical meaning until a supplement citing a 1939 reference book, Wright and Montag's Materia Medica (Pharmacology and Therapeutics, vol 10, 112); the same supplement's first reference to "iatrogenic" (medically-caused) illness dates from 1923. Edward Tenner argues that this is due to a change in the Western worldview and that before the 19th century, people saw all consequences as equal rather than "side" effects. Negative effects from a medicine were taken as proof that the desired changes in the body were taking place.

Sources:
Edward Tenner, Why Things Bite Back: Technology and The Revenge of Unintended Consequences New York: Alfred A. Knopf, 1996.
The Oxford English Dictionary entry for "side-effect".

"I'd like two butterflies..."
You want a side with that?
Uh... sure.... I'll take the propagative effects with extra ketchup

Usually, a side effect is what you get when you do something, then let it interact with the rest of the world.

Take for example a butterfly, flapping its wings. The side effects of this may well be a hurricane on the other side of the world.

A side effect of eggs are chickens, and a side effect of babies are people. That is, side effects are all relative - they change as you change your point of view. From one view, the side effect of internal combustion engines is motion. From the other, the side effect is exhaust.

In general computer science, a side effect of a routine is any change to data that is not in that routine's immediate scope.

For example, the routine "bar" in the following Pascal code has a side effect:

procedure foo;

var i: integer;

procedure bar (j: integer);

begin
i := i + j; { side effect }
end;

begin
i := 0;
bar (3);
writeln (i); { also a side effect }
end;

---

The C++ Standard defines a side effect as "changes in the state of the execution environment", specifically,

(1.9.7) Accessing an object designated by a volatile lvalue, modifying an object, calling a library I/O function, or calling a function that does any of the above.."

A side-effect also has a meaning in programming.

A procedure has side-effects if it modifies variables other than its inputs and/or outputs. This is bad 'cause it makes you code not entirely clear to the reader (unless it is heavily commented... ya, right. That happens...).

A plain old side-effect isn't too bad. You can usually hunt them down by looking at public/global variables. However, sometimes side-effects are used deliberately as a way of having two procedures talk to each other. This is really bad, and almost impossible to debug. I call it communication by side-effects, and I hate it.

If you read the list of common side effects of just about any psychiatric drug, you'll notice that they are almost always all physiological effects. Things like dry mouth, dizziness, sleepiness, etc. Very rarely are any cognitive or psychological side effects included, even though many have been documented extensively. Mental side effects such as mania, suicidal ideation, increased anxiety, and more rarely delusions and psychotic behaviour, are actually very common.

In very many cases, a certain drug will work great for a patient, while the others in its class will have a decidedly negative effect -- sometimes even worse than the symptoms which prompted the prescription in the first place. Since psychiatry is far from an exact science, it is often a process of elimination when looking for suitable medication(s) for an individual.

Prime examples of dangerous physiological and psychological side effects can be seen with fluoxetine and paroxetine (known commercially as Prozac and Paxil. These antidepressants are very strong inhibitors of the CYP 2D6 enzyme, but are also metabolized by the very same enzyme. This means that for people that are already genetically lacking a normal level of CYP 2D6 (about 7% of caucasians), the drug does not work the way it is supposed to, as it is not cleared out of the body correctly. If you or someone you know is considering taking either of these drugs, I would suggest you ask the prescribing physician for a test of the patient's enzyme levels first.

Sources:
http://www.womenshealthpc.com/3_99/pdf/177PharmRounds3_99.pdf
http://www.livejournal.com/users/vorpal/371244.html