Evidence - some thoughts from my favorite book.

[delphi_ote]

Godel Escher and Bach: An Eternal Golden Braid, by Douglas Hofstadter

I chose my career as a result of having read this book. It's clever, philosophical, educational, and beautiful on so many levels. Turning to almost any random page will fill your head with things to think about. Hofstadter definitely earned his Pulitzer. If you heed only one sentence I write on this forum make it the next one. Read this book.

Today, I randomly flipped to a page where Hofstadter discusses "The Nature of Evidence." As usual, his observations are very insightful and deep. They also seems highly relevant to our usual discussions. I wanted to share some passages with you and get some thoughts on them.

Concrete examples of evidence delemmas crop up in regard to many phenomena of fringe science. For instance, ESP often seems to manifest itself outside the laboratory, but when brought into the laboratory, it vanishes mysteriously. The standard scientific explanation for this is that ESP is a nonreal phenomenon which cannot stand up to rigorous scrutiny. Some (by no means all) believers in ESP have a peculiar way of fighting back, however. They say, "No, ESP is real; it simply goes away when one tries to observe it scientifically - it is contrary to the nature of a scientific worldview." This is an amazingly brazen technique, which we might call "kicking the problem upstairs." What that means is, instead of questioning the matter at hand, you call into doubt the theories belonging to a higher level of credibility. The believers in ESP insinuate that what is wrong is not their ideas, but the belief system of science. This is a pretty grandiose claim, and unless there is overwhelming evidence for it, one should be skeptical of it. But there we are again, talking about "overwhelming evedence" as if everyone agreed on what that means!

I think that sums up nicely our usual run-ins with pseudoscience and believers. Consistency of evidence and logic are less important to them than their beliefs, so they sacrifice the former when the latter is challenged.

Hofstadter gives a very cogent analysis of just why evidence is so important, and then goes on to give his own opinion on just what it's all about:

My feeling is that the process by which we decide what is valid or what is true is an art; and that it relies as deeply on a sense of beauty and simplicity as it does on rock-solid principles of logic or reasoning or anything else which can be objectively formalized. I am not saying either (1) truth is a chimera, or (2) human intelligence is in principle not programmable. I am saying (1) truth is too elusive for any human or any collection of humans to ever attain fully; and (2) Artificial Intelligence, when it reaches the level of human intelligence - or even if it surpasses it - will still be plaged by the problems of art, beauty, and simplicity, and will run up against these things constantly in its own search for knowledge and understanding.

That is what I think we all intuitively understand here. We are after the most accurate truth and understanding possible.

I found this last passage to be particularly interesting and relevant...

Science is often criticized as being too "Western" or dualistic" - that is, being permeated by the dichotomy between subject and object, or observer and obserevd. While it is true that up until this century, science was exclusively concerned with things which can be readily distinguished from their human observers - such as oxygen and carbon, light and heat, stars and plantets, acceleration and orbits, and so on - this phase of science was a necessary prelude to the more modern phase, in which life itself has come under investigation. Step by step, inexorably, "Western" science has moved toward investigation of the human mind - which is to say, of the observer. Artificial Intelligence research is the furthest step so far along that route. Before AI came along, there were two major previews of teh strange consequences of the mixing of subject and object in science. One was the revolution of quantum mechanics, whith its epistemological problems involving the interference of the observer with the obserevd. The other was the mixing of subject and object in metamathematics, beginning with Godel's Theorem and moving through all the other limitative Theorems we have discussed. Perhaps the next step after AI will be the self-application of science: science studying itself as an object. This is a different manner of mixing subject and object - perhaps an even more tangled one than that of humans studying their own minds.

By way, in passing, it is interesting to note that all results essentially dependent on the fusion of subject and object have been limitative results. In addition to the limitative Theorems, there is Heisenberg's uncertainty principle, which says that measuring one quantity renders impossible the simultaneous measurement of a related quantity. I don't know why all these results are limitative. Make of it what you will.

I think we see here very clearly why the "woos" rush after quantum mechanics, psychology, relativity, and genetics. These are fields which confuse our notions about the distinction between observer and observed. I think we also see very clearly here just why they are so utterly, hopelessly wrong to abandon science and logic. Science is the methodology by which we are coming to understand ourselves more deeply than ever before. The disciplined logic and self-criticizm of science is steadily increasing our understanding of the observer and observed.

Does anyone have any thoughts on the last paragraph about limitations? It's a very deep question. I think an answer would teach us a lot about the nature of observation.

 

[logical muse]

I'll just add this:

The computer programmers out there will get a kick out of the recursive stories.

It's my favourite book too.

 

[Z]

OK, I might just have to get a copy of this one.

Just those paragraphs alone were works of finest art... well, aside from the typos (I hope I can blame del?)

 

[Jeff Corey]

Does anyone have any thoughts on the last paragraph about limitations? It's a very deep question. I think an answer would teach us a lot about the nature of observation.

Problem with the word "limitive". I have a couple of good dictionaries and I tried a google. What does it mean?.
One problem I had with that book was his tendency to coin new words.
But Robert Rosenthal has explored the interaction between the experimenter's expectations and the resulting data in great detail.
And then there's:
Herr Schroedinger, does your cat bite?
Nein.
Nice pussy.
MEOrrr.
You said your cat doesn't bite!
It's not my cat.

I still like that book.

 

[Dymanic]

I chose my career as a result of having read this book. It's clever, philosophical, educational, and beautiful on so many levels. Turning to almost any random page will fill your head with things to think about. Hofstadter definitely earned his Pulitzer. If you heed only one sentence I write on this forum make it the next one. Read this book.

I heartily agree. I remember the day I innocently plucked that book off the library shelf for the first time. It was the name Escher that prompted me to reach for it in the first place, and the cover art immediately confirmed that I had found something interesting, but of course I had no idea what was in store for me. How could I? How could anyone who hasn't had a go at GEB possibly have any inkling? And how does one even begin to try to explain what GEB is about to someone who hasn't read it?

 

[delphi_ote]

OK, I might just have to get a copy of this one.

Just those paragraphs alone were works of finest art... well, aside from the typos (I hope I can blame del?)

Yea. Sorry about that. It was a lot of typing to do directly from the book!

Problem with the word "limitive".

Like that one. It should be "limitative." Hofstadter notes that our exploration of the interaction between observer and observed has so far only resulted in the discovery of limits to what we can learn.

 

[delphi_ote]

It was the name Escher that prompted me to reach for it in the first place

I saw some original Escher sketches at the art museum this weekend, and I've been grinning ever since. It was definitely the highlight of my year so far.

And how does one even begin to try to explain what GEB is about to someone who hasn't read it?

It's totally, absolutely impossible.

 

[SezMe]

Just to add my agreement. My copy is just about worn out. I like to open it at random places and read a few pages for instant brain stimulation.

 

[logical muse]

I saw some original Escher sketches at the art museum this weekend, and I've been grinning ever since. It was definitely the highlight of my year so far.

Check out my site eschertiles.com (sorry, can't make it a clickable link till I reach 15 posts)

The link labelled 'interaction' is where the fun is.

Hmm... I should get around to updating it one of these days...

 

[DarkMythril]

GEB is my favorite book as well. Particularly interesting is the book's materialistic view of consciousness. I believe Hofstadter is involved with CSICOP in some way.

 

[Dymanic]

logical muse:
Check out my site eschertiles.com

I like it! The instructions thingy doesn't quite seem to work, though. Could you post the instructions here?

DarkMythril:
I believe Hofstadter is involved with CSICOP in some way.

Hofstadter is a CSICOP fellow. (The list also includes Marvin Minsky, Richard Dawkins, Daniel Dennett, Francis Crick, Isaac Azimov, and many other very familiar names). Circulation of CSICOP's journal, Skeptical Inquirer increased sharply following Hofstadter's Metamagical Themas article in the February 1982 issue of Scientific American.

Another personal favorite of mine is his Fluid Concepts & Creative Analogies, the first book ever sold on Amazon.

 

[logical muse]

I like it! The instructions thingy doesn't quite seem to work, though. Could you post the instructions here?

Sorry 'bout that. Basically, just click the little arrows and numbers and things, or just click the Random button.

Message or e-mail me for more info as it's probably off-topic, and being new here, I'm still a little shy.

 

[Jeff Corey]

...Hofstadter is a CSICOP fellow. (The list also includes ... Isaac Azimov, .

Sorry squire, he is an ex-fellow. Along with Sagan and Skinner and others who are dearly missed.
(No parrot jokes, please.)

 

[Dymanic]

Sorry squire, he is an ex-fellow. Along with Sagan and Skinner and others who are dearly missed.
(No parrot jokes, please.)

Well, ok, details are important.

The late Isaac Azimov is not only listed among current and former CSICOP fellows, but was one of its founding members. (Hofstadter is still alive and well and teaching cognitive science (among other things) at Indiana University at Bloomington).

 

[Z]

Isaac Asimov, if you please - not Azimov.

Thank you.

 

[Dymanic]

zaayrdragon:
Isaac Asimov, if you please - not Azimov.

No excuse, sir. The worzt part about it iz, I think I've made that exact same miztake before. For some reason, I seem to be particularly prone to typoz that involve that sort of swap. What's even worse is the way the third finger on my right hand seems to have decided that a cute trick is to occasionally substitute a zero for an uppercase "O", especially when I'm writing code. Before I learned to look for it, this produced some really hard-to-find bugs.

logical muse:
Basically, just click the little arrows and numbers and things

Heh. I wish I had a nickel for every time I've given the same advice: "Just frob on the buttons awhile, you'll figure it out." (I did, too; just after posting that). If you ever do get around to updating it, I'm wondering how much work it would take to modify that thing so it could produce the sort of regular tessellations of the plane called "Parquet Deformations" done by students of William Huff, and described by Hofstadter in Metamagical Themas (a slight change to one or two of the variables which propagates across an array of vertices in a horizontal display, so the pattern gradually morphs into something completely different).
http://www.cgl.uwaterloo.ca/~csk/washington/tile/parquet.html
There must be freeware tools for this. I haven't found one, but it looks like you're ninety percent of the way there.

Message or e-mail me for more info as it's probably off-topic

Oh yeah, the topic. Well...

I think we see here very clearly why the "woos" rush after quantum mechanics, psychology, relativity, and genetics. These are fields which confuse our notions about the distinction between observer and observed. I think we also see very clearly here just why they are so utterly, hopelessly wrong to abandon science and logic.

But they don't abandon science and logic. They just get it wrong. They want their arguments to sound logical, but at the same time they want freedom from the rigors of logical formalism. They want their arguments to sound scientific, but they want to avoid doing the work that real science involves. They want their intuitive notions about things to recieve the same weight as empirical evidence, failing to grasp that logic is an alternative to intuition, and mysticism an alternative to science.

That having been said, there are some areas where intuition is the best thing we've got. Modern science is the product of some thousands of years of trial and error by humans, and, applied to certain types of problems, it serves far better than human intuition. Human intuition is the product of some millions of years of trial and error by evolution, and, applied to certain types of problems, it serves far better than science and logic.

Reverse-engineering human intuition is a problem which has stubbornly resisted the application of the methods of modern science and logic (for a quick example, scroll to the bottom of this page and have a look at the other threads the bot decided were 'similar' to this one). One of the main things I get from Hofstadter is that this is not only likely to continue to be the case for a very long time to come, but it may be a fool's errand to begin with.

 

[logical muse]

If you ever do get around to updating it, I'm wondering how much work it would take to modify that thing so it could produce the sort of regular tessellations of the plane called "Parquet Deformations" done by students of William Huff, and described by Hofstadter in Metamagical Themas (a slight change to one or two of the variables which propagates across an array of vertices in a horizontal display, so the pattern gradually morphs into something completely different).
http://www.cgl.uwaterloo.ca/~csk/washington/tile/parquet.html
There must be freeware tools for this. I haven't found one, but it looks like you're ninety percent of the way there.

I really do need to put some more work into it... Real soon now...