?

Log in

No account? Create an account
You, there, at the back. Wake up. - Lindsey Kuper [entries|archive|friends|userinfo]
Lindsey Kuper

[ website | composition.al ]
[ userinfo | livejournal userinfo ]
[ archive | journal archive ]

You, there, at the back. Wake up. [Jan. 27th, 2008|12:37 pm]
Lindsey Kuper

Over the wire this morning I see that John Walker, sitting in his Mountain Stronghold in Switzerland, has released new versions of midicsv and csvmidi, a couple of nerdy Perl utilities that convert standard MIDI (musical composition) files to and from Comma-Separated Value (CSV) text files, which can easily be manipulated by programs to perform various transformations on musical works.

In other startling news, any two systems that can be modeled with whole numbers can also model each other.

Left as an exercise for the reader: write a song about MIDI. Submit your work as a CSV. (Come on, people, it's 2008. I can't believe this doesn't exist yet. Get cracking.)

LinkReply

Comments:
[User Picture]From: oniugnip
2008-01-28 04:48 am (UTC)
... so we have an equivalence class of systems that can be modeled with whole numbers? What sorts of things can or can't be handled like this? All sorts of things are encoded as streams of bytes -- is this the right intuition to take to this?

(tell me all about this, please :) )
(Reply) (Thread)
[User Picture]From: lindseykuper
2008-01-28 07:54 am (UTC)

Yes, I think so, probably!

Okay, so, Gödel worked out a way of assigning unique integers to each symbol and well-formed string in PM's notation, such that string manipulations in PM could be modeled with numerical manipulations (you know, sophisticated stuff like, oh, addition and multiplication). It turned out that there were two kinds of integers in the world: those that coded for provable strings of PM, and those that didn't. (Hofstadter cutely calls the former the "prim numbers".) (Interesting aside: you can't build a machine that will tell you if any given number is prim or not, 'cuz halting problem.)

But "primness" turns out to be a purely number-theoretical property and therefore representable in PM! That means that you can write a (really frickin' long) string of PM symbols that will assert, "A certain integer g is not a prim number." But! If you are Kurt Gödel in 1931, you will have rigged it so that said integer g is in fact the integer that codes for that very (really frickin' long) string of symbols! Aha! Paradox! I've caught you in your web of deceit, Bertrand Russell! And stuff.

Anyway, to quote Hofstadter, this is all possible because "numerical patterns have the flexibility to mirror any other kind of pattern". So if "mirror" means "model", and "pattern" means "formal system", and if you can model a non-numerical formal system with numbers, then you can surely model a formal system that's already numerical with different numbers, right? And if you can create a mapping from non-numerical system to numerical system, then you ought to be able to create mappings between numerical systems, right?

And speaking of Wolfram, does his principle of computational equivalence say the same thing? (Except he might be full of crap.)

Also, I think I'm'a start saying "scale-invariant" instead of "fractal". It's more descriptive and less trendy.

Also, I want an umlaut in my name. Let's change our names. You know you do, too.

(Reply) (Parent) (Thread)
[User Picture]From: keystricken
2008-01-28 06:32 am (UTC)
In other startling news, any two systems that can be modeled with whole numbers can also model each other.

But will we like what we get?
(Reply) (Thread)
[User Picture]From: keystricken
2008-01-28 06:37 am (UTC)
In parallel (major) thirds with oniugnip, a question: are there links you can share? I've run a few searches and haven't come up with anything.
(Reply) (Parent) (Thread)
[User Picture]From: pmb
2008-01-28 06:59 am (UTC)
This is basically a restatement of Godel's theorem, where he proved that if you could model whole numbers, then you could model all known mathematics, including the statement "this statement cannot be shown correct using manipulations of whole numbers in a general mathematical system".
(Reply) (Parent) (Thread)
[User Picture]From: keystricken
2008-01-28 07:03 am (UTC)
Oh.
... but I thought you couldn't do that.
(Reply) (Parent) (Thread)
[User Picture]From: pmb
2008-01-28 07:18 am (UTC)
You can make the statement, but not prove its truth value. Or if you can, then the statement is false and your system just proved a false thing. Thus, every mathematical system sufficiently complicated to model the whole numbers will either contain unprovable truths or provably true falsehoods. The system that modern mathematicians use - Zermelo-Frankel set theory + the Axiom of Choice (aka ZFC) - contains unprovable truths.

It also contains things whose truth value is independent of the axioms, which is a slightly different condition. The most famous of these independent things is the Continuum Hypothesis (which comes in both normal and Grand forms), aka "How many real numbers are there, really?" The current drive is to find some "natural" or "intuitive" axiom that will imply an answer to the CH. Nobody knows any right now, so the Continuum Hypothesis sits around being of indeterminate truth value. As for me, as I advance in mathematical maturity I find it hard to even accept the Axiom of Choice, which apparently allies me with the crankiest of old mathematical codgers.
(Reply) (Parent) (Thread)
[User Picture]From: lindseykuper
2008-01-28 07:44 am (UTC)
I got here by thinking about Gödel's theorem, so, points for you, but I don't think it is quite a restatement of Gödel's theorem. I make no claims about being able to completely model whole numbers!
(Reply) (Parent) (Thread)
[User Picture]From: pmb
2008-01-28 07:57 am (UTC)
The system I'm talking about models them to the point of addition and multiplication, and that the numbers can get very big and that's okay. If you have that, and zero and one, then Godel has a trap for you.

If you have integers but cannot multiply or add them, or only have a limited number somehow, then your system is strange, but also may not necessarily contain a trap.
(Reply) (Parent) (Thread)
[User Picture]From: lindseykuper
2008-01-28 08:06 am (UTC)
Sure, yeah. But that's not really what I'm talking about. I'm not talking about modeling the whole numbers, I'm just claiming that if you have a system A and a system B and they can both be modeled with whole numbers, then you can also model system A in system B.
(Reply) (Parent) (Thread)
[User Picture]From: pmb
2008-01-28 08:11 am (UTC)
Looking back over what I said, I totally look like a pedantic jerk.
(Reply) (Parent) (Thread)
[User Picture]From: lindseykuper
2008-01-28 08:39 am (UTC)
'sokay. How'd your talk go -- did they like the drawing?
(Reply) (Parent) (Thread)
[User Picture]From: pmb
2008-01-28 08:55 am (UTC)
I actually am giving this lecture on Thursday. For some reason, the prof asked me to give the lecture a month before I would need to give it, and then asked for a picture, which made me think about how I was going to talk long before I needed to.

http://www.cs.uoregon.edu/classes/08W/cis170/
(Reply) (Parent) (Thread)
[User Picture]From: lindseykuper
2008-01-28 07:40 am (UTC)
But will we like what we get?

Oh, probably not. See, this is why I secretly want to be a mathematician. Mathematicians get to say, "I have found an completely dreadful way of doing X and thus proved that X is not impossible! Woo-hoo!" Then they hand it off to the computer scientists, whose job it is to find ways of doing X that don't make people want to claw their eyes out.
(Reply) (Parent) (Thread)
[User Picture]From: pmb
2008-01-28 07:58 am (UTC)
Math is "What", CS is "How".
(Reply) (Parent) (Thread)
[User Picture]From: keystricken
2008-01-28 07:32 pm (UTC)
Yeah. I mean, maybe I could turn a celery bread recipe into 12-tone music, but why would I want to?
(Reply) (Parent) (Thread)
[User Picture]From: lindseykuper
2008-01-28 07:37 pm (UTC)

how most twelve-tone music probably gets written

Because it's 2 a.m. and the assignment is due at 9 a.m.?

(Reply) (Parent) (Thread)
[User Picture]From: keystricken
2008-01-28 07:50 pm (UTC)

Re: how most twelve-tone music probably gets written

Hmm.

Bizarro-Monica is offline, but your theory has merit.
(Reply) (Parent) (Thread)