Is all information physical? Implications on things like Object Process Methodology

Systems Thinking
1

Continuing the discussion from Object Process Methodology:

OPM is:

“Based on a minimal universal ontology of stateful objects and processes that transform them, OPM can be used to formally specify the function, structure, and behavior of artificial and natural systems in a large variety of domains.” - https://en.wikipedia.org/wiki/Object_Process_Methodology

and

“In OPM, an object is a thing that exists, or might exist, physically or informatically. Objects are stateful—they may have states, such that at each point in time, the object is at one of its states or in transition between states. A process is a thing that transforms an object by creating or consuming it, or by changing its state.” - https://en.wikipedia.org/wiki/Object_Process_Methodology

What’s tripping me up is the OR between physically and informatically regarding the types of objects that can exist in OPM. It seems to imply that there can be non-physical information, non-physical objects. Does this mean OPM assumes dualism? I’m not yet decided on this topic myself, but I’m not sure a formal system like OPM can sit on a fence.

My belief seems to be that all information is embodied in some physical medium. What would someone model in OPM as being a purely informational object? Would this only work for hypothetical objects, or models themselves? But wouldn’t these hypothetical objects/models still technically only exist in the modeler’s mind (i.e. embodied in a brain)? Or do we skirt around that complexity by simplifying the systems of interest, and ignoring the bigger picture? Would a nondualist modeller just simply never use a purely informational object while using OPM? Also, this just made me think of Charles Sanders Peirce’s triadic model of semiotics where objects and the signs that represent them always exist in relation to things like minds, an interpretant.

I am interested in how @daviding has resolved these sorts of things in his mind, given he is the one who introduced me to OPM.

Am I just being too philosophical here? :stuck_out_tongue: Though, these questions seem to have been relevant to all the Natural sciences too… As I was just exploring my interest in all of this by browsing the web I found that the physical nature of information has touched many academic fields through history: Physics, Biology, Computation, and obviously information science itself.

First I found this paper: “Information is Physical” by Rolf Landauer

The PDF at the link above lead me to: Information: From Maxwell’s demon to Landauer’s eraser

There are a lot of thought experiments that centre around the second law of thermodynamics where entropy seems to represent information lost in a system.

“Maxwell’s demon is a thought experiment created by the physicist James Clerk Maxwell in 1867 in which he suggested how the second law of thermodynamics might hypothetically be violated.” - https://en.wikipedia.org/wiki/Maxwell’s_demon

This potential violation is called Negentropy by some:

“The concept and phrase “negative entropy” was introduced by Erwin Schrödinger in his 1944 popular-science book What is Life?” - https://en.wikipedia.org/wiki/Negentropy

Biological systems also seem to be based on information too:

“Both James D. Watson,[2] and Francis Crick, who jointly proposed the double helix structure of DNA based on X-ray diffraction experiments by Rosalind Franklin, credited Schrödinger’s book with presenting an early theoretical description of how the storage of genetic information would work, and each independently acknowledged the book as a source of inspiration for their initial researches.” - https://en.wikipedia.org/wiki/What_Is_Life%3F

There have been more recent experiments that have begun to test some of these older thought experiments, and the assertion by Landauer that “information is physical”. For example:

These articles seems to be less about refuting that information is physical, and more about showing that the second law can be violated… and that information can be used to reverse entropy and create work.

I wonder what David Hawk might think about all this. He seemed to use entropy and the second law a lot in his thinking about the arrow of time when I visited him at his home last summer.

Relating to systems thinking, I’d guess Ashby’s concept of variety fits in here somewhere.

2

Hi there,

I have a hard time following what all these people are discussing, let’s see where my thinking is flawed

When it comes to information systems (which are tasked to process information, and not doing some other things in which they secondarily produce/emit or “consume” information), information itself is always immaterial, virtual, and that’s per definition because we use the word to refer to abstract patterns of states, which can be - with the help of encoding schemes - be represented/captured/processed by non-existent virtual/imaginary machines/information-systems or actual physical ones.

It’s true that information can’t really (in reality) exist without being encoded in a physical carrier medium/system, but the same information can be encoded/expressed/transmitted/translated-into different physical media/systems, so the abstract pattern of hypothetical states is not bound to a single actual machine/system, which is why one may think about both being separate (they’re indeed in two different worlds, for which different rules apply - information doesn’t suffer from physical scarcity and only keeps multiplying, etc.). Now, every information processing machine (and other), being something that exists in the physical world, in its activities moves around matter and energy (without energy, things can’t move around, so state couldn’t change, so it could still be encoded information, but no processing), is observable from the outside and information about how it’s doing its information processing may too be captured (not exactly of what we usually call “metadata”, but roughly along these lines), and this could invite an attempt for trying to describe the abstract information as tied to the machine, the physical manifestation that happens to process it, but that’s hardly ever possible for the two reasons of timing and the other that one would need finer observation equipment than what the machine uses to do its processing, and if we’re already on some atomic level, that gets difficult to do on a large scale, and what’s observing/capturing the observation equipment, etc.

So in this regard, I would assume until convinced otherwise, that “information” (as the concept) itself is never physical, and it’s physical manifestations/encodings are a separate thing altogether, conceptually. For other systems that aren’t information processing systems, but perform some other primary actions, OK, you may observe them and describe what they do in abstract information, and then process such captured information with some other system (not with the observed one because that isn’t an information processing system), and here I don’t see how such an activity can lead to the question if information is physical or not. You’re already inadvertently creating an abstract state pattern and are transferring to a different physical carrier/host medium because you can hardly ever fully capture all the internal (!) operations of the first system under observation and it’s complex interplay with other systems as well as the many physical effects it’s involved with up to its place within the universe (therefore the capture is a simplification and abstraction, not accurately representing the real thing, and how could it for the immense amount of rather “meaningless” or potentially meaningful data, and what second system could then process it), which is why system modeling/thinking does describe an inside and an outside, with the internal details being abstracted away behind the system boundaries (reminder: these aren’t static, can change and overlap!), so it’s more or less unlikely to ever fully, accurately describe/encode the physical operations of a system in information. Abstract or manifest information is a series/pattern of states, and may inherently inadequate to represent the operations of a system, because within the system, a lot of different sources of state change do something in parallel, and the outside observer has just a few sensors, might not be entirely aware of the (changing) internal processes/setup of the system, or in other words: to accurately represent the system, it would need to be the system (also affects AI and stuff, trying to use formal electronical information processing to represent what’s going on in a biological information processing system, which are fundamentally different) and maybe at the same physical place at the same physical place too for there’s your entropy/decentralization.

So indeed we may only, exclusively encounter information manifested/encoded in its physical carrier medium (no surprise, ourselves currently being a physical system ourselves), but let’s not confuse the information itself with its encoding because the latter can change apart from the former, or why confuse energy with information (is it energy needed for changing a state or maintaining/storing a captured state, do we then care more about the pattern of states or how much energy is needed in an information system to maintain the captured/encoded/stored pattern?)?

I don’t understand Maxwell’s demon: introducing two rooms and the demon, doesn’t this change the setup? Obviously the universe isn’t exclusively governed by entropy, as outside of thermodynamics, such demon setups likely do exist, right?

Entertaining bonus: https://www.youtube.com/watch?v=Y0MsrrTo8jY#t=1h22m20s (from 1:22:20)

3

Now reading this Landauer article. It’s always problematic to look at the electronical, digital computer as the only or most important information processing system, it’s just that we know it to some extend because we’ve built it. Keep in mind that the physical current flowing through transistors and what not is used for a formal computational device, and High/Low is a more or less only arbitrarily used to model a minimal state change, from one state change to the other, to enable the most minimally required mechanics to enable an information processing system, it at least needs a means to record a state difference, at the very least. Sure, transistors with their low voltage requirement on their gate are helpful with it, but it could as well be a multi-state machine (fuzzy logic) or use other means than electronics (mechanical computation, brain, or some other information systems human-made or out in nature), doesn’t matter at all. In the computer, work-energy is needed to affect the state change, and we loose some between the transition from Low to High (only because we artificially decided on a threshold of what should be considered High and what Low, with an ineffective gap in it, and then, electronics are indeed analog, so the state change wastes energy inefficiently in the time gap of waiting for the transition between High and Low - Jaron Lanier for example too mentioned that this is why we emit the heat, unused energy wasted for no other reason than to maintain our formal, logical computation of arbitrarily so only two states, and now consider how accurately or efficient this can represent another real system using some other physics, but on the informational level, there’s no difference between the physical carrier medium).

Then the Zuse view that the universe may be one big computer, it’s probably the other way around, that all physical activity could be recorded as information and therefore processed, computed on. For information to be meaningful and not just data or noise, we would have to ask who’s doing this computation, for what purpose, and how would the information encoded/computed in the universe be interpreted by whomever is running it? Otherwise it may be data, but utterly meaningless to anybody.

Discussion of the Turing machine, read them as universal computational patterns/principles for logical operations, no need to be digital or even electronical, works on paper or with water or with/on anything else as well. But this amazing computational device for formal logic, the way it is doing it, is not the only way of computation, and therefore can hardly reflect/represent accurately or efficiently other forms/occurrences of computation.

I can’t see how erasing memory or logic operations “throw away” information or loose energy, that’s not what the information is and what the energy goes into. Sure, an information pattern can be lost, by creating a different one. We hardly can keep all newly created informational patterns captured for all of eternity, or is this what the article is asking for? Then, the energy isn’t the information, because the same computation can be performed on paper or with water and no electrical energy in our electronics sense would be lost, so the other forms of energy needed for performing the same logical operation under other circumstances isn’t related to the information, but for performing the state change (which then results in a different informational pattern, but assume that you know the desired pattern in advance, you could just directly encode/capture it without wasting the energy for computing it, so the energy in this particular case goes into the processing, not maintaining storage/state).

It seems the article goes on exclusively looking at electronic, digital computation, which is not the only and therefore in itself not necessarily much related to information processing. Whatever is described there for a more efficient computing device, the reader may simply assume that such devices potentially exist somewhere in nature (and if not, what if all information processing systems by their nature (of formal logical computation) have some waste/leakage, but because of their benefits, there’s a host system tasked to maintain the supply of energy to keep it operating), so we don’t learn much about information and other physical carrier media and computing systems from the deficiencies of our own digital information processing machines. If it’s about making our particular digital scheme more efficient in the use of electronic energy, OK, sure, why not.

Another quick thought: consider the natural state of matter and energy in the universe. We don’t necessarily consider all of it “information” because usually it’s noise/randomness to us, not meaningful in our perspective (apart from a full capture of physical state we’re far from understanding, observing and correctly interpreting). For matter, these patterns of state are important and meaningful. Now we want to change this default state into a different state that’s meaningful to us, for which we have to invest the energy to effect the change that naturally isn’t necessarily meaningful for the particles, energies, and the universe etc. One cool trick is to lock our patterns into the physical encoding of its carrier medium that’s also meaningful and inherently maintainable (stable) by the host system, and we get memory that doesn’t require energy to maintain its state. Systems that on their own perform changes based on energy sources they have available themselves, rather unlikely because what they want and do is usually different from what we want and do, so these state changes tend to be meaningful to either the one or the other system (but rarely for both), as we and them are different, separate, independent systems (not mentioning all the different things that go on at the same time at a sub-system level, which may interfere/conflict).

Another quick addition: Imagine an information processing system that happens to not be loaded with meaningful information. It can surely perform its operations by its physical design, but garbage in will result in garbage out (for how would it “enrich” the information with meaning or introduce the latter, as such would be relative to a different, separate system from itself it has no knowledge about nor insight into), so despite the physical machine as the carrier medium is perfectly fine and executing its actions including investing energy for state change, none of it is a useful statement about the nature of information.