Hacker News Reader

Comment by anigbrowl | Hacker News Reader

If the FISA decided that the NSA's activities were illegal, why would it keep rubber-stamping the warrants?

First, I don't agree that they rubber-stamp warrants, but believe they evaluate them on a case-by-case basis - not least because their decisions might be classified now, but may well be declassified in the future, and their reasoning will be held up to scrutiny.

More importantly, just because an agency is found to have operated illegally does not mean that everything the agency does must then be rejected. It's easy to find examples of police officers or even entire departments being corrupt, but it does not follow that everyone who is arrested by the police is necessarily innocent, or that all arrests are flawed.

Arguments of the kind you make above involve a fallacy of composition, one which seems to appear ever more frequently on HN. http://en.wikipedia.org/wiki/Fallacy_of_composition

View on HN

I'm not sure how anyone can has enough reasonable information to make a case that FISA doesn't rubber stamp cases. It seems to me it works that of the way large organizations navigate political nuances internally; people who don't run the technology and platforms make the decisions as it pertains to influence that, is often, good for self-preservation and fiscal motivation - the operators implement the systems as they see fit and generally don't have "real" oversight from the approvers. Even if they do - generally they don't have enough understanding to enforce it if they wanted to. It's just a matter of mincing words at that point.

I think until we see clear evidence from examples the number of cases denied are an obvious minority. Based on what we know from public disclosure, there is abuse. That to me states there is a good indication rubber-stamping has in fact occurred and should be reviewed. But, since it can't be...

The trustworthiness of the system is not there anymore and it has degraded to a point where most informed citizens would err on the side of saying there is something wrong. There are some grave issues that need to be addressed in short order or our government is going to push any data related industry the way of safer harbors.

"I'm not sure how anyone can has enough reasonable information to make a case that FISA doesn't rubber stamp cases."

The burden of proof is on the positive, not the negative.

In fairness that runs in both directions - we would certainly be better off with some proof of FISA courts' effectiveness, even if were historical (eg declassification of past decisions from >10 years ago, or the establishment of objective criteria for declassification).

It would also help significantly if the default for their opinions not be that they are automatically classified and put the burden of proof on the government in each individual case to justify why the public shouldn't know what the law is.

Well, opsec. Spying is by definition a clandestine activity, and I think there are good arguments in favor of spying - the main one being that while it might seem nicer not to spy on anyone, you end up being less informed, and uninformed executives are more likely to end up going to war, which is a lot more destructive than spying. If you read up on the Curveball story in Iraq, that's a classic case of what happens when you make decisions with inadequate information.

You don't have to disclose the specific targets you're spying on to publish the relevant court opinions. You can redact the names and still publish what the law is.

And there is a difference between collecting intelligence on foreign governments (which I don't think anyone could ever expect not to happen) and mass data collection. Part of the problem here is that spying on "terrorist" "suspects" is so broad and poorly defined that it could encompass nearly anything and becomes nothing but a convenient fig leaf for mass surveillance.

We lump too many things under the label of terrorism. Someone with the likes of a machete or a hand grenade is distinctly a law enforcement problem rather than a national security problem and cannot justify secret mass surveillance or data collection.

Once you take all of that out and get to the actual national security threats (i.e. nuclear or biological terrorism) you end up with a completely different threat model. More to the point, you need to keep a lot less secret because instances of attempted nuclear terrorism are much more rare. You don't have to classify every court opinion related to some idiot with defective instructions on how to make a pipe bomb just because you have to classify certain methods of preventing nuclear proliferation.

You don't have to disclose the specific targets you're spying on to publish the relevant court opinions. You can redact the names and still publish what the law is.

If you have to redact all the of the relevant facts, then any statement of the law is meaningless because it's unclear what sort of fact pattern it should be applied to. I realize this seems abstruse, but the fact pattern is very important in common law judgments. You can't just say 'the law is X' and have it be portable to any other case. The opinion would read something like: 'HELD: the government may xxxxxxx xxxxx when xxxxx xxxx xxx in a xxxxx and xxxxxx.' Read some judicial opinions in normal cases and then imagine how little sense they'd make if the facts were matters of national security and couldn't be published. For that matter it's quite hard for a lot of people to understand the law in many regular criminal cases where all the details are available.

Someone with the likes of a machete or a hand grenade is distinctly a law enforcement problem rather than a national security problem and cannot justify secret mass surveillance or data collection.

9/11 only involved boxcutters, but I'd say that it presented a rather significant national security problem when those turned out to be sufficient to hijack planes and fly them into high-value targets. There's an assumption in some corners that since we're now more aware of the risks, no hijacking can ever work again, but that's not the case.

You're also ignoring the fact that many kinds of low-level terrorist activity aren't open to direct investigation by law enforcement. For example, suppose you pick up signals about a plot involving Mr X, who has not yet arrived in the US but whose phone number you have managed to identify. If he's in a country that's friendly to the US you might be able to get their law enforcement people to investigate him, but if not then it's logical to monitor who he communicates with for prophylactic purposes.

>I realize this seems abstruse, but the fact pattern is very important in common law judgments.

You're basically assuming they would have to redact everything. Just replacing names and addresses with tokens that don't identify the specific subjects would get you most of the way there unless the specific fact pattern is unique to the suspect, and then you can redact what is necessary to make it less specific.

In addition to that, once any given investigation is over the opinions and redactions that were held secret for the purposes of that investigation should be published, and it shouldn't be so easy to keep things secret for decades just by making a facile claim of national security.

>9/11 only involved boxcutters, but I'd say that it presented a rather significant national security problem when those turned out to be sufficient to hijack planes and fly them into high-value targets.

9/11 was not a national security problem. 9/11 was a mass murder. It was shocking but we have inflicted more damage on ourselves in our overreaction to it than the terrorists did in committing it. More have died avenging the victims of 9/11 than died on 9/11.

A national security threat is a nuclear bomb, or a plague, or a foreign government infiltrating high level government offices in preparation for some kind of Communist takeover. Something that would be a factor of 1000 worse than 9/11 and present an actual threat to the security of the entire nation. This watering down of "national security threat" to mean any jerk who might kill some people with an IED is just a ruse to rationalize the use of extremist measures against everyone rather than taking into account the proportionality of the threat.

>You're also ignoring the fact that many kinds of low-level terrorist activity aren't open to direct investigation by law enforcement. For example, suppose you pick up signals about a plot involving Mr X, who has not yet arrived in the US but whose phone number you have managed to identify. If he's in a country that's friendly to the US you might be able to get their law enforcement people to investigate him, but if not then it's logical to monitor who he communicates with for prophylactic purposes.

So get a warrant and conduct surveillance on him then. That's not the same thing as spying and collecting data on everyone "for prophylactic purposes" at all.

That's true, though I'm not speaking to that point at all. Rather, I wanted to prevent a debate following from someone attempting to make a point that's technically indefensible but not necessarily sound. Expressing the same point without that kind of demand ("prove this isn't happening") would be fine.

But I do agree with you - this can be applied to FISA courts' efficacy. But that's not entirely a rebuttal since it's equally valid in both cases. Saying that in response as a rebuttal would wrongly imply it's fair to demand a proof of the negative just because it's valid for FISA effectiveness as well.

I agree. The government has never proven to us that FISA works, therefore it should be shut down until such proof is forthcoming. Remember that FISA was designed to solve a problem: rogue intelligence agency actions. If it has solved that problem, then we need proof, otherwise we must assume that the problem still exists.

This is true; I don't disagree.

The burden of proof applies to those who's liberties are infringed. If my rights are being stripped, the burden of proof should not be that I need to prove the NSA is infringing those liberties, it should be that they need to prove that they aren't. With National Security Letters, the recipients can't even acknowledge that they've received them, and with the FISA court, we have no way to know.

You missed the point.

In logic, it's fruitless to try to place the burden of proof on a negative statement. Example: "Prove there isn't an invisble ball of undetectable mass swirling in my hand."

What I was saying is a general rule that's valid across every subject in logical debate, and it is valid as a response for the sentence I quoted. What you're saying isn't a rebuttal to my point.

I have to take issue with the point about negative statements, surely it's no more fruitful to say "Prove there IS an invisible ball of undetectable mass swirling in my hand."

Wouldn't it be more accurate to say that it's fruitless to place the burden of proof on a statement which is syllogistically isolated from any statement known to be true?

No, it is more fruiful to say, "Prove there is..." because that's the basis of scientific/mathematical rigor. If you can prove it exists, it does. If you can't prove it exists, it doesn't, for all intents and purposes. You then act as though it doesn't, because until you can find evidence for it, it might as well be nonexistent, and you should act that way to be logical.

My point wasn't about proving a negative. More to the point there is no oversight so we don't know how bad it really is. I'm calling for some sort of protective measures and putting an end to this nonsense. The current administration is taking an ends justify the means attitude, not that they are probably the first to do so, but after being caught in lie after lie they aren't to be trusted either.

> In logic, it's fruitless to try to place the burden of proof on a negative statement.

That's unnecessary. We've already seen, via the leaks, that dragnet surveillance is happening. The NSA's claim that domestic communications are sacrosanct are put to lie by the revelations that the DEA and IRS are both using laundered NSA domestic surveillance data. We're now well beyond the need to prove that the government is in a state of sin with regard to the Constitution. The issue now is what remedies we'll pursue.

The problem: you can just rephrase the claim to have a positive statement. Now, If I'm to claim that the FISA is NOT effective in their role, and the burden of proof would be on you.

Yes, very true! But this isn't a problem because they're two separate things. The issue of rubber stamping and the issue of effectiveness are related but discrete problems to consider when talking about FISA courts' - whether or not they are effective does not necessarily confirm rubber stamping, though there could be a correlation.

But, yeah, I agree. It remains to be proven that FISA is effective.

"can has" -- And here I thought I proofread it. #fail

"First, I don't agree that they rubber-stamp warrants, but believe they evaluate them on a case-by-case basis"

The FISC is called a rubber stamp because they almost never reject requests for a warrant. The government can be almost certain that no matter how outrageous its request is, FISC will say "yes." That they may be evaluating requests on a case-by-case basis is irrelevant if they always come to the same conclusion.

> The FISC is called a rubber stamp because they almost never reject requests for a warrant.

Given that they initially provide feedback when they have problems rather than outright rejection, and that the government can, if it isn't willing or able to address the feedback, simply withdraw the application, its not entirely surprising that they don't issue rejections, per se, nor does that necessarily indicate that they are acting as a rubber stamp.

We'd probably get a better picture of whether they were a rubber stamp if they were required to give a rejection with comments if the application wasn't sufficient on its face, and we had statistics on that.

I disagree that this argument involves a fallacy. The following statements are not equivalent:

1. It seems like at that point it should deny every single request going forward until the illegal activities have stopped.

2. Every request going forward is illegitimate.

(2) is false, but I believe that (1) is plausible, especially if there are any difficulties for FISA to determine whether any particular NSA request is legitimate or not.

I think that they are; if you're sufficiently convinced of 1), then 2) follows as a matter of logic.

Not necessarily. The court can do (1) in order to punish the government for not complying, similar to excluding improperly obtained evidence that would otherwise be relevant.

Only if you define "legitimate" so that what you say is tautological; not if "legitimate" means, for example, "concerning a targeted and specific terrorist threat".

They don't rubber stamp warrants because there are no warrants. See the XKEYSCORE presentation. They just need a loose reason that the person may or may not be affiliated with a foreign entity by up to five degrees of separation.

Your argument is with the GP, not with me.

Reply on news.ycombinator.com