[Cryptography] More magical encryption thinking from James Comey

[Henry Baker]

At 10:44 AM 4/10/2016, Jerry Leichter wrote:
>> 'every time you hear somebody making a slippery slope argument, an alarm should go off in your head.  There is a reason your professors call this "slippery slope fallacy."   It could be that if you take one step you’ll inevitably fall down a slick slope, it could be.  It depends a lot on what kind of shoes you’re wearing, whether the slope is a stairs slope, and whether there’s a railing.'
>
>You're missing what he's getting at here (and, granted, he doesn't do a good job of explaining it).
>
>*As a general matter*, slippery-slope arguments in the legal context are consider very weak.  Legal principles and decisions have to do with human interactions and human realities, and in the real human world, clear bright lines rarely present themselves.
>
>When is someone legally responsible?  It's obvious that a 4-year-old should not be; it's obvious that a 30-year-old should be.  But where's the bright line?  We arbitrarily choose 18 (for most purposes).  Is there any real difference between 18 minus one day and 18?  Clearly not - but we have a rule, because there are circumstances in which we have to be able to decide.  And the whole point of "rule of law rather than rule of men" is *predictability*.
>
>And that's the way legal thinkers generally approach "slippery slope" arguments:  We know *this* should be regulated/made illegal/whatever; we know *that* should not.  We don't right now know exactly where the line should be drawn ... but that's the purpose of legal cases and precedent.  The Common Law has been crafting lines to divide this from that for hundreds of years.  That's the role of judges - especially those at the appeals level; it's also the job of legislatures (though they usually don't do nearly as good a job at it).
>
>BTW, reductio ad absurdum arguments get rejected even more strongly, and for similar reasons.  The view is "sure, you can take things to an extreme and make them look silly - but that's not the reality, the legal system doesn't do that."
>
>Now, I'm not trying to defend what Comey said, or even less to defend the his position on cryptography.  Nor do I claim that the legal system always gets this stuff right - clearly not.  I'm just trying to point out the kinds of arguments that get accepted in different spheres of thought vary, for very good reasons - and that making a list of "silly exceptional cases" that laws might reach will get you nowhere in the domains that really matter for *application* of laws.

I don't know much about the law, but it seems to me that a 'slippery slope' argument is essentially a Bayesian argument, which includes prejudices.

Unfortunately, Bayesian arguments take the blindfold off Lady Justice, so if I were being judged, I'd rather not have a biased judge.

On the other hand, public policy *must* take into account a priori probabilities, else we will spend billions of dollars trying to save a single life.  (Hence our insane focus on terrorists and guns instead of auto accidents and sugar.)

Medical treatments now include the extremely valuable "number needed to treat", which attempts to bring a Bayesian rationality to treatment plans -- including drugs.

So what "number needed to treat" is Comey suggesting?  That we destroy the privacy and security of a billion people in order to catch one bad guy?

Comey's proposal is not even remotely close to Blackstone's ratio:

"It is better that ten guilty persons escape than that one innocent suffer"

https://en.wikipedia.org/wiki/Blackstone%27s_formulation