"It's a book review -- doesn't really explain what the physicist is arguing in any meaningful sense."
Indeed, the review ends with this:
"Mr Vedral makes a persuasive argument for a third option: information can be created out of nothing."
That's a remarkable claim, and the review says zip about how it is presented. It seems that it would also be the key concept of the book, making this a pretty lame review.
Well it's not really remarkable at all, the argument is basically that information and entropy are mathematically equivalent and so information like entropy spontaneously increases.
Information and entropy are absolutely not equivalent; entropy describes the upper bound on information content, but says little about the actual information contained in a coding. The statement that information is spontaeously created isn't based in any physics I know of.
"The core of the argument (on page 201) I will not spoil here, but it's not wholly convincing on several grounds. First, although he does not further dwell on it, the author favors the 'Copenhagen' (or observer effect) interpretation of quantum physics, which is used to buttress & underpin his argument."
Oops. Or at least say I, not liking the Copenhagen interpretation at all. (Perhaps the most frightening thing to me about quantum mechanics is that I've found the many worlds interpenetration to be the best ... then again, in this area I'm a chemist, so these issues are "above my pay grade".)
Whether a physicist (in modern times and below a certain age) espouses MWI is quickly becoming a great litmus test for whether you should bother listening to them on anything that isn't established textbook material. Copenhagen is pseudo-scientific nonsense, and I'm amazed anyone can work on developing quantum computers while simultaneously denying the existence of what they're designing. I'd cut him some slack if this were just some tangential side point, but if it's critical to his argument the whole of it isn't worth anyone's time.
"Whether a physicist (in modern times and below a certain age) espouses MWI is quickly becoming a great litmus test for whether you should bother listening to them on anything that isn't established textbook material."
Errr, do you mean you should bother listening or not bother?
My understanding of MWI doesn't go much beyond what I could understand of an early Wheeler paper (I thought it was from the mid-40s, but Wikipedia says MWI starts with the thesis of a student of his with the first publication in 1957 ... and I read that paper (whenever it was published) 25 years ago so it's gotten pretty fuzzy).
Hmmm, Wikipedia says MWI "along with the other decoherence interpretations and the Copenhagen interpretation" are the mainstream interpretations, so I suspect that answers my question.
My understanding of MWI doesn't go much beyond what I could understand of an early Wheeler paper (I thought it was from the mid-40s, but Wikipedia says MWI starts with the thesis of a student of his with the first publication in 1957 ... and I read that paper (whenever it was published) 25 years ago so it's gotten pretty fuzzy).
Well, in this universe it started in '57. In yours, it seems, it is different. Where are you from traveller ?
The reason people can get away with Copenhagen is that decoherence is absurdly effective in all realistic cases, causing there to very little uncertainty about whether and when a measurement happened. Insofar as we can all agree when a measurement happened, physicists operating under Copenhagen are going to be just as effective as those who understand MWI. And for most of the theoretical and especially computational aspects of quantum computers, this is the case. The only place where someone operating under the Copenhagen interpretation is going to falter is when they have to think hard and in novel ways about how to prevent decoherence.
So I really don't think interpretation is going to be correlated with advances in the computational aspects (i.e., what kind of problems are quantum computers good for once they are built, etc.) but I would predict that future insights in how to build computers which are robust against decoherence will be dominates by those who understand MWI.
"Insofar as we can all agree when a measurement happened, physicists operating under Copenhagen are going to be just as effective as those who understand MWI."
Agreed. If you're only worrying about the mundane medium-sized world we live in and what QM predicts we'll see in it, MWI vs. Copenhagen won't matter. Here, it does, as he's apparently using Cop. as a part of his argument.
On the other hand MWI super-advocate David Deutsch says: "An engaging, non-technical exploration of what the new theory of quantum information and computation tells us about life, the universe, and everything."
No surprise here... Edwin Jaynes called this the mind projection fallacy whereas Ayn Rand called it the primacy of consciousness, an epistemological error. If everything is "information" then what are the physicists studying? To quote AR "A consciousness conscious of nothing but itself is a contradiction in terms". Full quote here;
http://aynrandlexicon.com/lexicon/consciousness.html
Consider a computer program with access to the memory in which the program is running and no other inputs. The program is information, and only has access to information. The information is embodied in some set of hardware, but the computer program has no way of knowing or accessing this hardware (and we assume the hardware is 100% reliable, e.g. no memory faults due to overheating RAM or whatnot). So from the perspective of the program, everything is information, including the program itself. Substitute 'conscious mind' for 'program' and you have a conscious mind in a world of nothing but information. The fact that a 'hard' world exists outside of the 'soft' world is irrelevant, by virtue of being unknowable.
This article seems to imply that the author equates Shannon entropy with information. Is that so? It has always struck me that Shannon entropy is a good measure of information, but is not a good definition of information itself. Like the difference between a vector and the magnitude of a vector, information and its measure are not the same. What is information itself? If this author actually has a definition of that, I would like to see it.
"You should call it entropy, for two reasons. In the first place your uncertainty function has been used in statistical mechanics under that name, so it already has a name. In the second place, and more important, no one really knows what entropy really is, so in a debate you will always have the advantage."
The big problem with Shannon entropy is that it's measured with respect to a probability space, so it's not an "absolute" measure of information. Kolmogorov complexity "solves" this problem by defining information content of a string as the length of the shortest program that generates the string. (You still have to pick a reference UTM, but this only affects the program lengths at worst by a constant.)
"Unusually for a physicist, Mr Vedral spends a fair bit of time talking about religious views, such as how God created the universe. He asks whether something can come out of nothing . . . Mr Vedral makes a persuasive argument for a third option: information can be created out of nothing."
Somehow this reminded be about this verse in the Bible:
"In the beginning was the Word, and the Word was with God, and the Word was God", http://bible.cc/john/1-1.htm
.. and the heavenly spirits decided to set the Word free with a GNU license, and released Open Office, and after fixing most of the bugs God saw that it was good (or at least usable).
Mistaking the deepest abstraction you are capable of for the fundament of all reality is hubris.
Philosophy is one of those subjects that most intelligent persons think they can master pretty well. It all seems pretty simple and straightforward. Especially physicists fall into that trap again and again, up to great string theorists putting forward anthropocentric theories of the universe. These people usually make trivial mistakes that any real philosopher spots in no time. Intelligence is not a substitute for knowledge and this guy seem to have fallen into the same trap as well. In this case you don't even need to be a philosopher. He defines information based on the frequency of 'events'. This means that both time and 'events' are more primal than his notion of 'information'. Consequently, 'information' is not at the root of anything. And how could it be, if we need minds and language to discuss 'information' in the first place?
He does not define information. He relates to the general accepted definition of information by Shannon. There is a whole research field on this area, called information theory.
The base of this definition of information has nothing to do with time. It is rather a mathematical measurement for entropy.
"He defines information based on the frequency of 'events'. This means that both time and 'events' are more primal than his notion of 'information'."
True, but what do you expect for a popular review of a book that itself is a popular review of some pretty abstract ideas.
The basic concept is not just that our universe can be described by statistical measures, but that the foundational laws seem to be governed by statistics. At the small scale (individual particles), quantum mechanics basically says that the universe runs itself by rolling dice in very particular ways. Events involving these particles seem to simply be the tangible embodiment of Shannon entropy. Even at the everyday scale of refrigerators and gasoline engines, entropy calculations are the bread and butter of thermodynamics engineers.
It is a major open question in physics whether entropy and statistics are truly foundational, or whether they are simply the large scale average result of a fine-grained deterministic foundation. The "Copenhagen interpretation" takes statistics to be fundamental, and by gosh the calculations sure work nice, but there is no evidence that it reflects a higher truth.
Re. the anthropic principle in physics, it is one approach to answering why certain laws of physics seem to be balanced on a knife edge between chaos on one side and stagnation on the other. If we posit that universes can be created with nearly arbitrary physical laws, then naturally we find ourselves in one that is neither too chaotic nor too stagnant for our existence. The creation of such a universe might be unlikely, but given enough opportunity it is inevitable. This allows for bizarre physical laws without requiring magic or divine intervention. (There are even evolutionary approaches, where events like black hole formation create a new universe similar to its parent. That way if you find yourself in a part of the multiverse conducive to your kind of life, neighboring universes are likely to be very rich in friendly conditions.)
"And how could it be, if we need minds and language to discuss 'information' in the first place?"
If the evolutionary approach were true, universes of pure thought could potentially be created, where the very atoms were symbols grounded in some reality. A mind there wouldn't pick up a pen to draw a symbol, it would just think it into existence. Or such a parent universe could create a universe like ours, to test whether symbol processing is a necessary law of physics, or if it can spontaneously arise from semi-chaotic fundamental laws.
Well put. The one alteration I'd make is that universes of pure thought would require a Tegmark Level IV type multiverse, not just a MWI or black-hole-spawning type multiverse.
I don't tend to find these varieties of ontological reductionism that compelling, though it is interesting to read when people propose new ones. Sure, you can construct a coherent account of how everything really is just X for a lot of X, and everything else is merely higher-level structure on top of it. But picking the X always seems to be a bit arbitrary. Here he argues for the primacy of information; traditional materialists argued for atoms; some dynamicists argue for dynamics and state spaces; phenomenologists argue for conscious experience; Cartesian dualists argue for both atoms plus consciousness (everything reduces to those two, but neither reduces to the other); social constructionists argue for human social relations; some mathematicians argue for atoms plus math; etc.
Although I don't like everything else they write, I sort of like the manifesto of the Object-Oriented Ontologists, which is more or less that, while objects like televisions undoubtedly are made up of atoms, can be analyzed using information theory, are phenomenological entities, have a (perhaps socially constructed) role in human society etc., those are each just methods of investigating televisions, and televisions themselves "really" do exist as well: http://ooo.gatech.edu/
I doubt it. Maybe if he said "monetization the root of everything".
Most philosophy of science arguments remind me of old costume dramas. They tend to say more about the times of the people who made them than about the times of the people depicted in the story.
Okay, I'm just going to come out and say it, since it comes to mind every time I see an article like this. Let me know if I'm off track.
Isn't it more accurate to say that "description," rather than "information," is the root of everything? Information itself is pretty abstract, so it seems far more useful to do it this way.
What is information really? Does information exist without a conscious being to interpret it? Are the laws that govern information simply laws that are created by conscious beings for the purpose of interpreting the information? Most laws that we are concerned with are derived from nature, through our senses and experience of cause and effect. Our senses allow us to perceive the changes in state of our surrounding environment. These changes in state is information.
I guess the point I am finally making, is that the statement above concerning information is expected in lieu of our current focus, reliance, and addiction on information as a representation of reality. I am not sure what the difference is between the laws that govern matter and the laws that govern information. We're still working with information in either case. We have to take measurements, interpret the data, make predictions, and build systems to exploit the "laws" we've uncovered.
I think you're substituting some colloquial fuzzy/subjective definition of the word "information" here as "a conscious being knowing something." But in the context of the linked article, we're talking about Information in the context of Information Theory. It is not subjective. The remarkable thing about Shannon's information theory is that he does formulate real laws of information on the same solid, objective footing as laws of physics.
But then, I could have stopped reading at your first line: it sums up your entire post.
It seems quite strange if, as alleged 3rd hand in this thread, Vedral is obsessed with quantum computing, but sticks with the antiquated and more or less content-free Copenhagen interpretation. MWI seems to me to be critical to quantum computing -- otherwise, where is all that extra computing power coming from?
That doesn't make as much sense as it sounds like it does.
From a macro point of view the multiple worlds are all identical. All. The business with hitler doing something different in a different world is nonsense, quantum uncertainty almost never has any effect in the macro world.
So when you are computing in your world, everyone is doing the same computation in all the worlds.
Part of the MWI is that each world has a slightly different version of the event, and you just see yours, which was effected by theirs. But in a quantum computer you specifically set it up so that your get a single right answer - but if you get just one answer, all the worlds are the same, so how do you get multiple worlds each working on it?
I did my Masters thesis informational ontology. Here's the thing: Shannon's information theory is a mathematical formalism, and is not enough to define information. Furthermore, information can only be defined within bounds of a model, and there are typically many ways and many abstraction levels available to describe a phenomenon, so the question of objectivity of information is still open.
If to assert objectivity you try to pick the lowest possible layer of abstraction at the quantum mechanics level, you lose the whole layer of semantic information, i.e. information about something, which is actually the primary concept of information, and "information" simply becomes a name for a set of mathematical formulas, and as such it may not be treated as a substance.
39 comments
[ 2.9 ms ] story [ 91.9 ms ] threadIndeed, the review ends with this:
"Mr Vedral makes a persuasive argument for a third option: information can be created out of nothing."
That's a remarkable claim, and the review says zip about how it is presented. It seems that it would also be the key concept of the book, making this a pretty lame review.
;-)
"The core of the argument (on page 201) I will not spoil here, but it's not wholly convincing on several grounds. First, although he does not further dwell on it, the author favors the 'Copenhagen' (or observer effect) interpretation of quantum physics, which is used to buttress & underpin his argument."
Oops. Or at least say I, not liking the Copenhagen interpretation at all. (Perhaps the most frightening thing to me about quantum mechanics is that I've found the many worlds interpenetration to be the best ... then again, in this area I'm a chemist, so these issues are "above my pay grade".)
Errr, do you mean you should bother listening or not bother?
My understanding of MWI doesn't go much beyond what I could understand of an early Wheeler paper (I thought it was from the mid-40s, but Wikipedia says MWI starts with the thesis of a student of his with the first publication in 1957 ... and I read that paper (whenever it was published) 25 years ago so it's gotten pretty fuzzy).
Hmmm, Wikipedia says MWI "along with the other decoherence interpretations and the Copenhagen interpretation" are the mainstream interpretations, so I suspect that answers my question.
So I really don't think interpretation is going to be correlated with advances in the computational aspects (i.e., what kind of problems are quantum computers good for once they are built, etc.) but I would predict that future insights in how to build computers which are robust against decoherence will be dominates by those who understand MWI.
Agreed. If you're only worrying about the mundane medium-sized world we live in and what QM predicts we'll see in it, MWI vs. Copenhagen won't matter. Here, it does, as he's apparently using Cop. as a part of his argument.
John von Neumann to Claude Shannon
The big problem with Shannon entropy is that it's measured with respect to a probability space, so it's not an "absolute" measure of information. Kolmogorov complexity "solves" this problem by defining information content of a string as the length of the shortest program that generates the string. (You still have to pick a reference UTM, but this only affects the program lengths at worst by a constant.)
[Disclaimer: Hobbyist, not expert.]
Somehow this reminded be about this verse in the Bible: "In the beginning was the Word, and the Word was with God, and the Word was God", http://bible.cc/john/1-1.htm
Philosophy is one of those subjects that most intelligent persons think they can master pretty well. It all seems pretty simple and straightforward. Especially physicists fall into that trap again and again, up to great string theorists putting forward anthropocentric theories of the universe. These people usually make trivial mistakes that any real philosopher spots in no time. Intelligence is not a substitute for knowledge and this guy seem to have fallen into the same trap as well. In this case you don't even need to be a philosopher. He defines information based on the frequency of 'events'. This means that both time and 'events' are more primal than his notion of 'information'. Consequently, 'information' is not at the root of anything. And how could it be, if we need minds and language to discuss 'information' in the first place?
What's important is coming up with a theory that can make accurate predictions. No need to worry about whether it actually corresponds to reality.
The base of this definition of information has nothing to do with time. It is rather a mathematical measurement for entropy.
http://en.wikipedia.org/wiki/Information_theory
True, but what do you expect for a popular review of a book that itself is a popular review of some pretty abstract ideas.
The basic concept is not just that our universe can be described by statistical measures, but that the foundational laws seem to be governed by statistics. At the small scale (individual particles), quantum mechanics basically says that the universe runs itself by rolling dice in very particular ways. Events involving these particles seem to simply be the tangible embodiment of Shannon entropy. Even at the everyday scale of refrigerators and gasoline engines, entropy calculations are the bread and butter of thermodynamics engineers.
It is a major open question in physics whether entropy and statistics are truly foundational, or whether they are simply the large scale average result of a fine-grained deterministic foundation. The "Copenhagen interpretation" takes statistics to be fundamental, and by gosh the calculations sure work nice, but there is no evidence that it reflects a higher truth.
Re. the anthropic principle in physics, it is one approach to answering why certain laws of physics seem to be balanced on a knife edge between chaos on one side and stagnation on the other. If we posit that universes can be created with nearly arbitrary physical laws, then naturally we find ourselves in one that is neither too chaotic nor too stagnant for our existence. The creation of such a universe might be unlikely, but given enough opportunity it is inevitable. This allows for bizarre physical laws without requiring magic or divine intervention. (There are even evolutionary approaches, where events like black hole formation create a new universe similar to its parent. That way if you find yourself in a part of the multiverse conducive to your kind of life, neighboring universes are likely to be very rich in friendly conditions.)
"And how could it be, if we need minds and language to discuss 'information' in the first place?"
If the evolutionary approach were true, universes of pure thought could potentially be created, where the very atoms were symbols grounded in some reality. A mind there wouldn't pick up a pen to draw a symbol, it would just think it into existence. Or such a parent universe could create a universe like ours, to test whether symbol processing is a necessary law of physics, or if it can spontaneously arise from semi-chaotic fundamental laws.
Although I don't like everything else they write, I sort of like the manifesto of the Object-Oriented Ontologists, which is more or less that, while objects like televisions undoubtedly are made up of atoms, can be analyzed using information theory, are phenomenological entities, have a (perhaps socially constructed) role in human society etc., those are each just methods of investigating televisions, and televisions themselves "really" do exist as well: http://ooo.gatech.edu/
Most philosophy of science arguments remind me of old costume dramas. They tend to say more about the times of the people who made them than about the times of the people depicted in the story.
Isn't it more accurate to say that "description," rather than "information," is the root of everything? Information itself is pretty abstract, so it seems far more useful to do it this way.
What is information really? Does information exist without a conscious being to interpret it? Are the laws that govern information simply laws that are created by conscious beings for the purpose of interpreting the information? Most laws that we are concerned with are derived from nature, through our senses and experience of cause and effect. Our senses allow us to perceive the changes in state of our surrounding environment. These changes in state is information.
I guess the point I am finally making, is that the statement above concerning information is expected in lieu of our current focus, reliance, and addiction on information as a representation of reality. I am not sure what the difference is between the laws that govern matter and the laws that govern information. We're still working with information in either case. We have to take measurements, interpret the data, make predictions, and build systems to exploit the "laws" we've uncovered.
But then, I could have stopped reading at your first line: it sums up your entire post.
From a macro point of view the multiple worlds are all identical. All. The business with hitler doing something different in a different world is nonsense, quantum uncertainty almost never has any effect in the macro world.
So when you are computing in your world, everyone is doing the same computation in all the worlds.
Part of the MWI is that each world has a slightly different version of the event, and you just see yours, which was effected by theirs. But in a quantum computer you specifically set it up so that your get a single right answer - but if you get just one answer, all the worlds are the same, so how do you get multiple worlds each working on it?
If to assert objectivity you try to pick the lowest possible layer of abstraction at the quantum mechanics level, you lose the whole layer of semantic information, i.e. information about something, which is actually the primary concept of information, and "information" simply becomes a name for a set of mathematical formulas, and as such it may not be treated as a substance.