The term "slippery slope" is a metaphor. Like all metaphors, it is a lie; someone making a slippery-slope argument is not literally sliding down a hill. Metaphors often illuminate a general pattern, but they can do so at the expense of obscuring the specifics.
The Federalist Society sponsored a talk by Professor Eugene Volokh on April 9, 2009 entitled "Mechanisms of the Slippery Slope," where Professor Volokh tried to fill in the specifics surrounding the metaphor. The talk is based on an article in the Harvard Law Review, and a short version was adapted for Legal Affairs. Commentary was supplied by the always voluble Professor Richard Epstein.
A "slippery slope" occurs when taking action A increases the likelihood of action B occurring. To make a convincing argument, the slippery-slope proponent must specify the mechanism by which the slippery slope operates (hence his title). One mechanism is changing costs. Suppose that action A is gun registration and action B is gun confiscation. The creation of a gun registry eliminates some costs of confiscating guns: it locates the guns, reducing search costs; and it creates probable cause for a search, reducing legal costs. Lowering costs hardly makes gun confiscation a foregone conclusion, but gun registration would alter the landscape.
The previous example assumed no change in attitudes, but attitudinal changes form a second mechanism. The Supreme Court in 1979 in Smith v Maryland decided that using a pen register was not a search within the Fourth Amendment. (A "pen register" tracks the time and length of a person's phone calls and to and from which phone numbers the calls go.) After 9/11, law enforcement wanted to access without a warrant people's email header information, that is, the sender, recipient, and timestamp. The public debate primarily featured two arguments: pro, that header information was just like pen register information; and con, that header information was much more revealing than pen register information. Missing, however, was any argument that Smith was wrongly decided. A common heuristic is that if something is legal, it is correct; thus, an argument that the Supreme Court was wrong was less likely to gain traction. Debate over action B proceeded with the assumption that action A is acceptable.
Many more mechanisms exist. A third is political momentum. Legislators and many voters care about the strength of interest groups but have no way to directly measure that strength, so they use the success of a group as a proxy for its strength. A victory can sway decisionmakers by persuading them of a group's power, making the subsequent victory more (or less) likely. A fourth form is small-change tolerance, where decisionmakers agree with actions A, B, C, and D if taken individually but would disagree if all are taken at once. This form is often described with the hackneyed example of a frog not noticing that it is being slowly boiled.
Knowing that an argument employs a slippery slope is not enough to evaluate that argument, but slippery slopes often represent realistic concerns. Of course, if we say that slippery-slope arguments are acceptable, everyone will want to make them.