"It from bit" or "Bit from it"

bohm2

I've always had some difficulties understanding the whole concept of information:

Is matter really just information? Does one mean just our information about matter? Isn't all information embodied in some more basic "physical" stuff? Doesn't information require something to be informed? Can information be "active" (as per Bohm/Hiley's proposed "informational field" that guide the particle)? Nevertheless, I thought this would be an interesting poll. So, if you had to choose, which of the 3 options do you find more reasonable:

1. Information is more primitive/fundamental than matter/physical/energy ('it from bit')
2. Matter/physical is more primitive/fundamental than information ('bit from it')
3. None of the above (e.g. neither is more primitive/fundamental than the other or the question is meaningless)

Some interesting quotes for both positions:

I. IT FROM BIT:
Introducing the Computable Universe
http://lanl.arxiv.org/pdf/1206.0376.pdf
The physics of downward causation
http://www.oxfordscholarship.com/vie...4318-chapter-2

II. BIT FROM IT:
Bit from It
http://platonia.com/bit_from_it.pdf
Against ‘measurement’
http://duende.uoregon.edu/~hsu/blogfiles/bell.pdf

Ken G

Given my other posts, you will no doubt not be surprised that I conclude that both "it" and "bit" rely on each other to exist. I think Wheeler is right on target that the universe is participatory, and that physics is about information, but we must also recognize that what we mean by "the universe" is not actually the universe "out there", it is precisely the universe that we interact with. We know about the universe by interacting with it, and so our every language about physical reality is actually a language about our own participation in physics. The universe is not participatory, we are, and we are the physicists too.

Thus to object to information being housed in our minds, rather than in the universe itself, is to miss that the two are not so easily separated after all. Yet there should need to be some "it" for the "bit" to exist, while there needs to be "bits" for us to know anything about "it", so both require the other to mean anything-- or else neither term belongs in any language. In the immortal words of the Talking Heads: "that word does not exist in any language, it will never be uttered by a human mouth." Hence, neither "it" nor "bit" is more fundamental-- they each give the other its meaning, and if you push either one too far to their extreme, they disintegrate into the other. It is the place where they interact that the concepts bear fruit.

audioloop

"IT"

Gold Barz

What does Wheeler's "it from bit" say about consciousness or subjective experience?

bohm2

At times I kinda get confused with these arguments. But my guess, is that Wheeler would probably say that physics doesn't have much to say on subjectivity? Consider this quote by him:

Law without law
http://what-buddha-said.net/library/...ithout_law.pdf

But then again, I'm not sure? I have a lot of trouble understanding some of their positions. There seems to be this dualism that really confuses me: bit (information) vs it (matter), mind vs body, configuration space vs 3-space, etc. What makes it even more difficult is there seems to be this implicit assumption that we really have a clear conception of what we mean by "matter" or "information" is. But I'm not sure we do?

Paulibus

I think this confession from children's literature just about sums up how little is known (including by the erudite Wheeler) about the fancied link between information and physical reality. To clarify what is supposed, I'd like to have explained exactly how the information in the quote, lost if all our books were to fall through an event horizon around a collapsing mass concentration, would nevertheless remain smeared out, as it were, all over the horizon. Are material binary bits scattered like confetti all over the horizon? And if they're material, why don't they also fall through it?

ConradDJ

I posted some thoughts about this a few years ago... currently working on a paper that talks about how information needs to be defined, in physics. The main thing is that if we define information quantitatively, in terms of “bits”, we can conveniently bypass the quantum measurement problem – but end up without a clear connection between information and “matter” or any other aspect of physics.

For me the basic question is how information actually does get defined, in the physical world. We know it does, because we can see it and communicate it. But we don’t yet have a good framework for understanding how measurement and communication actually work.

This is from a draft of my current paper –

Despite all we know about physical interaction at the quantum level, we still conceive these basic information-processes essentially the same way Plato and Aristotle did. We talk about information as if it were built into the intrinsic nature of things in themselves, as if it were inherently well-defined, apart from any context. In this framework, to measure something essentially means to copy data out of the object onto something else, like making an impression of the object on a wax tablet, or recording its length with a ruler. To observe a thing means to make a copy of it in your mind.

But at a fundamental level, there are no physical processes corresponding to this notion of how information works. Making copies of given data is something that hardly happens in the physical world, though living organisms have found ways to do it, as the basis for their evolution. And human culture depends on our being able to reproduce information, in many different ways.

But the basic information-processes in physics are quite different. They're not about replicating information that's already objectively pre-defined, in itself. Instead, they're about setting up contexts of interaction between particular local systems that can physically define the information in the first place.

chiro

Regardless of the model for causality and regardless of the actual constraints used whether they are implicitly or explicitly defined, language is still required not only for the ability to represent and describe something, but also to compute.

Computers work with information and the structure must be known to the actual processing unit in order for a computation to take place.

This requires essentially one to describe the linguistic representation of the information being processed and thus indirectly a natural decomposition in the context of the computation exists.

Everything is completely based on a descriptive capacity, and the structure of the language itself as well as the entire spectrum of decomposition tells the scope of the analysis and subsequent computation that may be performed on the information.

Decomposition is the basis for analysis. It doesn't matter if the language is a spoken one, an aural one, a written one, a mathematical one: there has to exist methods to decompose things process them in some kind of decomposed state and then synthesize the results to bring something new.

For a programmer that is used to dealing with myriads of data structures all day, this is not really something out of the ordinary especially if the structures themselves are complicated.

The thing is though that computations at the fundamental level need to deal with some kind of representation of some sort and this requires a descriptional capacity.

cobalt124

It from bit. Damn, I've got it the wrong way round. Bit from it. It is more fundamental. I don't see how it could be any other way. Why would bit be more fundamental? Why would it be both? The only reason I can see is to give a better way of describing how the universe works, not what it fundamentally is. I don't see what previous posts are saying about actual reality. If it's just about usefulness of concepts then I don't see a big issue.

Could someone change my it from bit vote to a bit from it vote please?

Paulibus

Chiro, you have it exactly right: language is the key requirement because, as you put it, a "descriptional capacity" is needed --- I hope you mean to define and deal with information (and also its converse, entropy).

Well, it just so happens that such a capacity is innate in human beings, to a degree that far exceeds that in our fellow animals. Courtesy, somehow, of Evolution; we don't know exactly how and when we were lucky enough to evolve this capacity, but we now have it, in spades. Hence the inordinate length of threads in this forum about such matters as the philosophy of that most arcane, elegant and useful language, mathematics.

And indeed, hence physics and these forums, in which our attempted quantitative description of the contingent physical circumstances we find ourseves in is discussed. Not to be taken as engraved in stone as discovered truth, since the same information about circumstances may often be adequately described by us talkative primates in several distict dialects.

chiro

The important thing, building on Paulibus' comment is that descriptions are relative and not in a vacuum.

Different descriptions of something have relativity with other descriptions in the way of the similarities and differences with respect both the universal set of all descriptions, as well as with what something has been related to. So if we look at A compared to B of which both are in U, then A can not be just related to B, but also to U as well.

It is a subtle point, but when the notion of relativity is lost (i.e. the intent to always compare possible descriptions together for some kind of utility and to ascertain the advantages and disadvantages of one over the other), then that is the situation that creates a vacuum and all point of reference (which requires relativity is lost).

In the context of the computation, computational models are also relative to one another just like descriptions of information are, which is an important to consideration for constructing any argument or any discourse over why one computational model is preferable or not preferable to another.

Paulibus

Again, I agree that: as a practical matter for driving our understanding deeper, in physics as well as in writing computer software.

But it is also important to keep in mind that relative descriptions only describe and inform; they do not uncover what is. Indeed, I suspect that we have become overconfident about the use of the dangerous words "are" and "is", as well as about the deep importance of information, which is easily coloured by the fallible perspectives of folk who ferret information out of our complicated world.

Such overconfidence may well be derived from the simplicity, apparent truth and everyday utility of the information in, say; "this rose is red" — a statement that conveys quite different information when unpacked from the contrasting perspectives of a poet or a physicist.

Even the perspectives of respected Billy Thomson, who maintained that "atoms are vortices in the luminiferous ether", and of those who have more recently proposed that "atoms are tiny strings" may mislead if this information doesn’t lead to predicted phenomena that are later confirmed.

Information as a deep aspect of reality seems to me an idea that can be taken too seriously, especially in physics, where the Baconian tradition of making predictions and confirming them seems both necessary and sufficient.

I think "it from bit" is nonsense. How can such a claim be tested?

chiro

Anyone can measure something: that isn't really hard at all. You can ask anyone to observe what is happening and if you need an instrument to do it, and the analysis needs to be carefully done, then so be it.

But without any kind of relative analysis, it means absolutely nothing.

If I give you a string of random digits in an unknown language and you look at it, then it will probably mean absolutely nothing to you. If it's in a language you understand, then it will be decodable, but still won't contain anything that's useful.

But when the language has context, then it's an entirely different matter. When it has context it has a relative aspect to it, and the relative aspect means that the description can be differentiated from other descriptions that are otherwise also realizable but not: in short the context is a product of having a point of reference.

The way that measurement done now has no context.

If you want to give it context, the thing that needs to be done is to consider the alternatives and then consider any discussion as to why one should be realizable or preferred over another.

You ask how can it be tested? Well the thing is that you are forced to work under uncertainty: it is the foundation of science. We assume uncertainty and do the best we can.

The other thing that you have to realize is that most of science is inductive: we take results and try to wildly extrapolate general principles in most cases from said results, observations, and measurements.

One needs more than simple tests: you can measure all you want, but without a solid argument or any comprehension, the whole point is completely moot.

The linguistic argument is simple: in order to analyze something you need a descriptive capacity. Without a descriptive capacity you can not analyze: without analysis, you can't compute. Without computation, nothing can be transformed, or evolve.

The argument is based on our current theory and definition for a computer and subsequent ideas are found in computer science and linguistics.

It makes sense at least from some point and does not require measurement of any kind to understand.

If you want to test this in the lab, construct a statistical test to do so and perform the test on your data: that's how you can test the assumption. First come up with a null hypothesis and a normal hypothesis, differentiate the region for acceptance or failure to accept the null hypothesis and publish your findings. That's it.

But despite this, it would be interesting to ask you why the above assertion is wrong in any way whether partial or non-partial.

Whovian

As the device I'm on can't load PDFs, I can't read the computible universe argument. However, I'm fairly certain Quantum Mechanics' randomness prevents the Universe from being computed, as it's been proven that software algorithms can't generate any sort of randomness.

Whovian

(Again, I can't edit posts on this flippin' thing, but I wish to make a correction. I didn't mean to call the Universe computible, but rather, the properties of it and its contents.)

chiro

The flow-control need not be bound through deterministic paradigms: the flow-control can be based on non-classical means.

The argument relates to the representation of information having to have some kind of discrete representation and this being a requirement for a computation under paradigm to actually manifest.

Again, the issues is about requiring a descriptional capacity for analysis of any sort, but the requirement of a Turing machine or deterministic equivalent to underly the computational model for the universe.

Paulibus

Chiro: What "above assertion" are you talking about? the one that "is wrong in any way whether partial or non-partial"? I can't see anything wrong or even controversial in what you posted!