[hobs]

I think the big difference between the 1700s and today that seems unacknowledged in your post is that most communication happened face to face and was not documented for the government to intercept.

You and your wife could not be forced to testify against each other(Spousal privilege), but your private sms conversation absolutely could be. (What was possible as only a private conversation is now easily sent across the world, and as a consequence is sniffed and stored by potentially many parties.)

This is not a surprise for anyone who understands that they are sent in plain text, but from the context of people, those conversations would still be considered private communications between spouses.

I am not saying the law gives a shit about the distinction, but persons absolutely do. I hope you now understand the sentiment a bit more.

[tptacek]

I agree with this! I think the law is going to need to change, too.

The problem is, no matter how we change the law, the law is going to demand a balance of interests. We can outlaw coerced production of personal communications between relatives and friends, but there are still going to be communications the public has a right to access.

The technology is not going balance those interests. It's morally neutral, and incapable of concerning itself with the needs of society to stop organized criminal enterprises.

Again: for me, the wins of strong crypto clearly beat out the losses. But it's not insane that people --- especially people who don't work in technology --- might weight the factors differently.

[Grishnakh]

Answer me this, because I really don't know:

Back in the 1700s, if you wrote a private letter (on paper/parchment, with a feather ink pen as they had back then), in a foreign language, and the court system wanted to use this as evidence, could the government compel you to translate it?

Or, suppose you developed your own cipher (they had ciphers back then, I'm sure). Could the government compel you to decipher the message back then?

The use of encryption really isn't much different from this.

[csandreasen]

I don't think that's an appropriate analogy. You may as well just say "In the 1700s, if you wrote an enciphered letter..." since cryptography isn't a new concept. Various schemes have been used to protect military and diplomatic communications for centuries. If you did so then or now, you wouldn't be under any obligation to reveal the contents, but you take on the additional burden of actually performing all of the necessary calculations, securely destroying the scratch paper you used in the process of encrypting the message, handling key management and distribution, securing the areas where the encrypting/decrypting is taking place (you wouldn't want the redcoats barging in the hour or so while you're in the middle of converting the plaintext to ciphertext), etc.

Nobody does that anymore. You're instead using a tool that someone else made, and either that tool or the other person is handling all of the hard work. Tools definitely can be regulated - I need a license to drive; I need to register my car; I need to go through a background check to own a gun; I can own a gun, but if I misuse it I go to jail; felons can't purchase guns legally; I can't buy a nuclear weapon or the fissile material needed to make own.

The questions that policymakers are fumbling through right now are things like "how (if at all) do we regulate tools and the companies that make/distribute them if those tools allow people evade law enforcement?"

[hobs]

I think you are not wrong, but the point I was trying to make is that the whispered conversation of romantic partners is now something that is recorded and written and sent, the reality is that the law didnt change to become more draconian, the people's behavior changed so that the law FEELS more draconian.

Basically, when most of your society is illiterate, they are not creating evidence for LEO to subpoena in the first place.

[pc2g4d]

But with backdoored encrypted communications the government would have instant access to ALL enciphered letters at once, regardless of the seeming guilt or innocence of the sender or recipient.

To me that seems like the key difference. If all but face-to-face communications are electronic, and no electronic communications can be strongly encrypted, then the private sphere is greatly reduced and many things once considered private become public.

But on the other side of things, with strong crypto many things once considered public would become private.

There doesn't seem to be an easy way around this choice.

[tptacek]

Hold on. With status quo electronic communications in the 80s and most of the 90s, the government retained instant access to communications for investigative purposes, and no serious objections were raised --- just as nobody objects to the idea that the police, searching your house with a warrant, get instant access to letters you've left on your desk.

Instantaneity can't be the fulcrum of this debate, because it's been the norm since the beginning of English common law.

There has to be some other principle at stake that can argue against decryption backdoors. And I think there are such principles! But I think it's important that they be articulated carefully.

[pc2g4d]

Yes, in previous decades the government could access electronic communications under the third party doctrine[1] that says the fourth amendment only applied to "papers" held by the individual / in their home, not to communications voluntarily placed in the hands of a third party. On the other hand, far less of people's lives was conducted online. I expect that's why no real objections were raised.

Because much intimate communication has moved from in-person/on-telephone/other-instantaneous communications to asynchronous platforms hosted by third parties, in effect much that was "private" is now effectively "public". What once required a warrant now does not. Many things previously subject to protections against unreasonable searches are no longer so protected.

[1]: http://www.abajournal.com/magazine/article/the_data_question...