Court Accepting a Slippery Slope Argument:

There's an interesting example in the recent Google subpoena ruling. The issue in this part of the opinion was whether the subpoena was likely to jeopardize Google's trade secrets; the government said it wouldn't, because the government would agree to keep the revealed information confidential; Google argued otherwise, especially because there'd also be follow-up subpoenas, so the chances of some unauthorized disclosure would become even greater.

The court generally accepted Google's argument, expressly using a slippery-slope-related metaphor (the "thin end of the wedge"):

Google's remaining trade secret argument is that despite the narrowness of the sample provided, it would become entangled in the underlying litigation where further discovery would risk trade secret disclosure.... On the one hand, a determination of the propriety of further discovery is for another set of motions, and not the one presently before the Court. On the other hand, further discovery in this case that would require disclosure of Google's trade secrets is not merely a remote possibility....

In light of the comments of Plaintiffs' counsel at the hearing, the Court can foresee further entanglement based on Plaintiffs' challenge to the Government's ultimate study. In litigation where the ultimate question is ... fundamentally about limiting the access by minors to [Internet] adult material, it is quite likely that Plaintiffs will challenge the sample produced by Google as not representative of what minors search for or encounter on the Internet. Such an inquiry would require additional discovery, some of which may implicate Google's confidential commercial information.... [T]his Court is concerned that a narrow sample of Google's proprietary index and query log, while in itself not likely to lead to the disclosure of confidential information, may act as the thin blade of the wedge in exposing Google to potential disclosure of its confidential information.

This is an example, I think, of what I called a multi-peaked preferences slippery slope, but perhaps should have called an "unstable compromise slippery slope." Right now (at the position I call 0), the bottom of the slope (which I call B) — broad discovery of Google's data by a wide range of litigants, not just the government but the plaintiffs in the underlying cyberporn law case — is unlikely. But if the court allows the first step (which I call A), namely discovery by the government, then for various legal reasons (such as the felt need for both sides to be able to impeach each other's evidence) step B would become much likelier. Therefore, in evaluating whether the 0->A move is sound, the court shouldn't just focus on the costs and benefits of A, but should also include the possibility that 0->A would lead to A->B.

Camel (A) sticks his nose under the tent (B), which collapses, driving the thin end of the wedge (C) to cause monkey to open floodgates (D), letting water flow down the slippery slope (E) to irrigate acorn (F) which grows into oak (G). [Illustration by Eric Kim, from Eugene Volokh's idea.]

ed in texas (mail):
Is the effect the same if the monkey doesn't wear glasses?
3.20.2006 1:14pm
Rube Goldberg = genius
3.20.2006 1:26pm
JosephSlater (mail):
I admire Eugene's work on this issue generally, but the visual/graphic here is absolutely brilliant.
3.20.2006 1:37pm
Jay C (mail):
You forgot one thing, though: if there is any problem with any of the intervening steps, one may end up with
(H) Fruit of the poison tree
(insert small skull-and-crossbones where appropriate)
3.20.2006 1:47pm
Richard Bellamy (mail):
The alternative, of course, is to permit the camel to stick his nose under the tent only after the court orders the monkey to move his chair.
3.20.2006 1:55pm
jnet (mail):
stunning display of rhetorical tropes.
3.20.2006 2:48pm
Dilan Esper (mail) (www):
I think the "unstable compromise" is one of the most potent of slippery slope arguments. A nice legal example of this is the express limitation in Griswold v. Connecticut, limiting the right to use contraception to married couples and premising it on the privacy of the marital bedroom.

Anyone could see Eisenstadt v. Baird, which extended the right to single people, coming. A contraception right that was afforded only to married couples was an unstable compromise, because there would be strong arguments about not treating single people equally, and also about contraception being even more important for singles who are probably somewhat less likely to want children.

Of course, in that case, I don't mind the end result; I don't think singles and married couples should be treated differently with respect to the ability to purchase and use contraceptives. But if you DO mind the end result, the fact that the particular proposal is likely to result in an unstable compromise can be a very powerful argument against the proposal.
3.20.2006 3:15pm
BU2L (mail):
Completely off topic but - Prof. Barnett, any chance we can get a Sopranos discussion thread?
3.20.2006 3:31pm
How far from the tree did the acorn fall?
3.20.2006 3:34pm
Kovarsky (mail):
That's just an odd slippery slope argument to make, since surely once google can prove the data sample is representative within some margin of error then more discovery to increase sample size N would be an overbroad subpoena. There would be no slippery slope if the court were familiar with stats....
3.20.2006 4:43pm
Zubon (mail) (www):
Isn't the camel's nose supposed to go in the tent?
3.20.2006 5:20pm
Michael B (mail):
Somewhat OT, but it's difficult to have much sympathy with the enormously wealthy and powerful at Google - or Yahoo! - if the China related reports are true, i.e., that they're not only kowtowing to the Chinese People's Central Committee's demand that all types of social/political censorship be performed (such as pertains to their gulag's, extreme forms of persecution, other sundry topics) but are even turning in "offenders" to the Central Committee and subsidiary authorities.
3.20.2006 7:09pm
lucia (mail) (www):

I suspect the governments complaint would be that the type of of searches in represented in the sample were qualitatively different from those used by children.

Sample size would then be irrelevant, because collecting 10N times as much qualitatively inappropriate data does not improve the answer to the question you want answered. It improves the statistical precision of the answer to some entirely different question.
3.21.2006 12:25am
SenatorX (mail):
That essay on slippery slopes EV linked is brilliant. A study of Causality really and if I understand correctly an argument in favor of conservatism in action.
3.23.2006 10:47pm