[FRIAM] hidden

[uǝlƃ ☣]

You jumped close to where I was about to go! Now that we have some conception of how this principle is holographic (everything's there on the surface, all we need is the way to read it), I'd like to demonstrate that we don't *need* "interiority" to argue for privacy. But my argument differs a bit from yours below. Yours below argues that the keys/ways to read the surface may be inaccessible. My argument is that there are *many* ways to read the surface, some of which may even be mutually exclusive [†]. Further, I think there's a no-go result lurking beneath that we might get to if we get past the lower order results.

I'll start with comprehension of strings to work my way to the simplest form of privacy. Stronger forms might follow. Given the string "tin", what ways are there to transform the string? And by what ontology do we decide which of those transforms produce something meaningful? Obviously, an English speaker would land upon the reverse() function, reverse(tin) => "nit". A programmer might use cons(cdr(),car()) => "int". Someone who triggers on "interiority" might use the simpler cdr() => "in". >8^D A chemist include selecting just the first 2 to get Ti. We could elide the middle to get TN, Tennessee. Etc.

The idea is that when a *surface* presents itself, what are all the possible ways to *decode* that data? And, further, which decoding processes produce meaningful results? (I'd argue this is the definition of intrusion detection, anti-virus software, code breaking, etc.) If a super simple example like the string "tin" shows an explosion of possible transformations, what can we get from a more realistic example like finding Waldo in a kid's book? Or (Satan help me) interpreting an ink blot?

Given that combinatorial explosion, it is practically infeasible [‡] to slice/rebundle the possibly meaningful transforms down into a collection that can be handled in any small amount of time/resources. Hence we get the simplest form of privacy: "privacy through obscurity". None of us will ever know Frank's image of some childhood friend because there are simply too many ways to parse the data. David Icke can always recant some silly conspiracy theory by saying "that's not what I meant". Trump can avoid responsibility by claiming he said something sarcastically. Etc. There's no way for us to know, for sure, that a chosen decoder isn't the wrong decoder.

Of course, this raises the question of big data, AI, Moore's Law, etc. With enough time/resources, we can brute force our way through it. With enough crafty logic, we can winnow the space down. So, if anyone cares, we can take further steps to establish higher order privacy. Note that I'm *still* assuming that everything's there on the surface. I'm trying to use the position I infer from EricC and Nick to *demonstrate* privacy.


[†] I've lazily made this argument a lot by referring to Rosen's defn of complexity or von Neumann's extrapolation of Gödel, Wolpert's limits of info, etc. But nobody seems to acknowledge those and/or show me how I'm wrong about them. [sigh]

[‡] And perhaps impossible in principle.

On 5/18/20 6:46 PM, Jon Zingale wrote:
> I will attempt to switch sides and argue for why his mind may be private.
> 
> Firstly, while we may only need to know some combination of
> /transformations/ which will allow us to know his mind, it may
> be the case that those transformations are not accessible to
> us. As an example and in analogy to computation, it may be the
> case that we are not the kind of machines which can recognize
> the language produced by a mind. While we as observers are
> able to finite automata our way along observations of Frank,
> his mind is producing context-free sentences, say. I don't
> entirely buy this argument, but it also may be defendable.
> As another example/analogy, we may be attempting to solve
> a problem analogous to those geometric problems of Greek
> antiquity††. It may take a psychological analog to Galois theory
> before we understand exactly why we can't know Frank's mind.

-- 
☣ uǝlƃ