Quantum mermaids
Published by admin October 14th, 2007 in Doctors of PhilosophyI was watching this video of Roger Penrose at Perimeter Institute giving a lecture called Are we due for a new revolution in fundamental physics? This is the kitschy slides interpretation of Quantum Mechanics. One of his slides is this Quantum Mermaid:
The title reads: A new angle on the basic paradox of quantum mechanics. Mermaid is the connection between Classical World and Quantum World of physics.
Penrose does not realize that his peddling of mermaids as personification of Quantum Paradox puts him in the company of other promoters of real mermaids such as P. T. Barnum.1
Mermaids are not paradoxes. Mermaids are hoaxes.2 Anyone who exhibits mermaids as paradox3 is a hoaxer.
Mermaids of Quantum Mechanics are by-products of legal physics. Laws of physics are absolute by definition. Therefore, the only thing left for physicists to do is to invent mermaids to fit observations into physics doctrine.
If physicists were scientists they would reason like scientists:
“Guys, we are observing a mermaid. There must be something wrong with our assumptions, laws, our set up, our interpretation or all of the above. Let’s go back to the drawing board.”
But physicists are professional doctors bound by the legal doctrine. Physicists will observe mermaids rather than doubt physics dogmas.
As discussed here Feynman eliminated one of the Quantum mermaids called Uncertainty Mermaid by introducing new notation. So, Feynman proves that Quantum mermaids are not paradoxes but artifacts of physics canon that can be eliminated by using proper notation.3595 ringtone free nokianokia 1260 ringtonenokia 3390 ringtonesringtones 3595 free nokia6600 ringtone nokia polyphonic freeringtones absolutely verizon freeringtones music 100 charge free noget cent ringtone 50 know wanna Map
Mermaids are not paradoxes. Mermaids are hoaxes.
Hm, I always thought they were Walruses.
There doesn’t seem to be any paradox or hoax associated with walrus.
I agree that Ideas are not property.
Interesting looking article. I’ll read it.
Pioneer, thanks!
(Be warned that I think that particular article has some errors when it comes to distinguishing between something that is “quantum” and something that is just probabilistic. Better would be a version where the state probabilities are modeled as actual qubits/bloch spheres. I am working to redo it like that for a future post, but I am still learning this stuff so it is taking awhile…)
Coin, I read this article and the previous one. I find your research very interesting for the reason that I want to understand why this kind of original research could not yield valuable insight into what you are researching, namely, microscopic phenomena? I believe that what is thought in physics textbooks is a language developed in the 18th century and it is stuck there. It is not clear that physics formalism must be used exclusively rather than modern methods.
Hi Pioneer1,
Dunno if I’m posting this too late for you to see it
but
The reason why I do not think that the quantum cellular automata stuff I was trying to work with are likely to be practically useful for modeling quantum phenomena, is because the classical equivalents (which are very well understood) have yet to be shown useful for modeling classical phenomena.
There are people who try– there is work out there where people try to model fluid dynamics, for example, by means of cellular automata systems mean emergent behavior is similar to that of actual fluid systems. And this guy named Stephen Wolfram actually did write a popular book, called A New Kind of Science, mostly about trying to use iterative systems like the cellular automata, instead of equations, to model various systems that arise in science (although to my understanding he wound up saying a lot about computation theory very little about any other science). However it seems to me very unconvincing at the current time that cellular automata are a generally useful tool for producing models of physical systems.
This said, classical cellular automata often can tell us a lot of interesting things about computation theory. And computation theory is right now something to which understanding the difference between classical and quantum behavior is quite relevant, since we are very interested in to what extent quantum computers are more powerful than normal computers! So one of the things that interests me about the quantum cellular automata stuff is whether it could potentially provide interesting models for telling us things about the limits of quantum computers. I know there has been some work already on quantum cellular automata but I have not had the chance to really explore it yet.
Also, all this said, one of the main reasons why (I think) cellular automata have not so far been useful for physical modelling is that CAs are inherently discrete structures, whereas almost all interesting physical systems have some kind of continuous behavior. However, it may be there are interesting systems in some new physical theories which contain legitimately discrete structures, like spin networks. It seems much more likely that cellular automata might actually be legitimately relevant to or connected to these in some way. I would be very curious about this possibility but I don’t think I’m qualified given what I know right now to comment on whether that’s realistic or not…
Do you think modelling by cellular automata is rule based or data based or something else, as I tried to discuss here
Why contrast cellular automata with equations? Equations used in academic physics are tools to manipulate ossified physics doctrines about quantum interpretation of the world invented by European physicists born in the 19th century.
New methods of looking at this type of phenomena, even if they ultimately fail, will be beneficial, I believe.
I object to the classification of systems as physical and non-physical. “Physical” is a word invented by Newtonian physicists to make the world Newtonian. There are no physical systems. Internet is a same kind of system as planetary system.
I believe that cellular automata will model systems that can be modelled by iterative tools. (As you mention below re: spin networks.) There is no one absolute way to model observations. For instance, both commutative and non-commutative methods are valid in certain realms.
Sounds good. But I know no more about computation theory than what I read in Wikipedia. It seems like a fundamental field.
I question anything labeled “classical” by physicists. Physics is a scholastic corporation based on a hierarchical bureaucracy. Hierarchical bureaucracies are held together by a sacred code, the faith of the organism, that constantly grows by adding new code to it.
Every new generation adds his own code and labels what came previously classical. The legal code cannot be removed because without it the organism would disappear. Practitioners retire the legacy code gracefully by labelling it classical. This is the reason why physicists still use methods invented in the 13th century.
And because they sanctified those old methods, new ways of looking at the phenomena like you are attempting to do are frowned upon.
Observations are discreet but what is observed is continuous. That’s why I would like to understand what physicists mean by quantum phenomena. If they mean absolute indivisible particles, that’s wrong. Absolute discontinuity is not observed in nature.
There is also the role of probability in measurements in the quantum realm.
If the criteria is probability there is no difference between quantum and classical realms. In both scales measurement is probabilistic. This is easy to see: there is no measurement without error. The error is the probability.
So to me, as a person totally ignorant of QM, it appears that the difference is only in the degree of probability. Because of the scale or precision issues the measurements beyond a certain level appear ruled by probability while, say, measuring with a telescope, does not appear to be probabilistic.
Thanks for your nice comments.
Hi, sorry I took awhile to get back to you but:
If I understand the definitions you give there, whether cellular automata are rule based or data based depends on what we mean by “cellular automata”. There are two ways of using cellular automata to compute something.
One is the case where the set of cellular automata rules themselves perform the computation. For example there is a 1D cellular automata ruleset in Wolfram’s book which “computes” the prime numbers because whenever you place a single cell within this automata, that cell eventually grows into a complete Eratosthenes’ sieve. This kind of “cellular automata” is usually not what people talk about when they discuss cellular automata, because they’re not as interesting; but they are “rule based” in a very pure sense.
The other way of doing a computation by “cellular automata” is when you assume some rule set to be given, and then you construct patterns targeting that rule set. When you do this, you design a pattern which, when it evolves according to the ruleset, will interact with itself so as to produce some result you want. This kind of CA is what people working on cellular automata are usually interested in. These CAs are “data based” in a very, very pure sense, since the rules in this case are both predetermined and chosen to be very very simple and so the pattern specification (the “data”) does all of the “work”.
The reason why the words “cellular automata” could mean either of these things is that technically a “cellular automata” refers to the pair of these two things– a set of rules, and a starting state (a “pattern”) together. In concept you could think that this suggests a third option for constructing cellular automata which is both “rule based” and “data based”, where some of the complexity of the final computation comes from the rules chosen, and some of the complexity comes from the starting state chosen. But people don’t work with CAs that way very often.
Overall though the thing that is interesting to me about CAs in this context– although this has very little relevance to physics!– is CAs give us a very clear separation between what is “rules” and what is “data”. Because there is this separation, when you are working with CA is possible to make the data simple and concentrate on the rules, or vice versa. Compare this with a programming language or even a turing machine, where the rules and the data get all mixed up and it is more ambiguous which is which (since in those cases, data directly encodes rules). This makes CAs useful for discussion of what it means for example for a computation to be “universal”, which is I think a big part of why Wolfram is so interested in them.
Well, when I contrasted cellular automata with equations above, I think I was maybe using poor terminology. “Equations” and cellular automata are just two different ways of describing a computation, or a relation. However when we say “equations” we usually are limiting what we are talking about quite strictly, because there are only certain things we use equations to describe. When I mentioned someone using “equations” to describe nature, what I really mean is they’re describing nature by processes which can be described using only polynomials and sinusoids and series and the operations it is traditional to find in an equation– I don’t know what the best way to describe it is, but there is some sense in which traditional math limits itself by the form of what can be expressed using an equation. At the least, if something is simple enough to be described by an “equation” I think it must be what computation theory calls “primitive recursive”! Traditional equations impose a certain kind of simplicity, and that kind of simplicity is not metaphorical or aesthetic, it represents an actual barrier in a computation-theory sense. When I contrasted equations with CAs I really was trying to contrast CAs with the certain kind of simplicity that equations impose.
Cellular automata however are universal. It is very easy to construct a cellular automata which is as powerful as a Dell computer (but not as fast), in that if you give it the right input it can do ANYTHING. CAs don’t have to be primitive recursive. They can be really, really powerful. That’s bad. I don’t think scientists want their models to be powerful, they want them to be simple. They want the models to be simple because they believe, or hope, that nature is simple; and they want the models to be simple because if they are simple they will be easy to analyze! CAs are very difficult to analyze because you can’t make assumptions– if you have a CA that is supposed to model wind patterns or something it’s very hard to generalize and say well after 1000 steps the wind will be doing this and this, because you’ve given the model too much power and so for all you know any given input could be simulating a Dell computer. Now, maybe we need to give our models more power than we give them now, I don’t know. But if we give them too much power then it will become almost impossible to say anything about how they behave…
I’m not sure if what I just said made sense or not
Well, when I said “classical” above I really just meant “not quantum”. But…
I very much share your frustration at not being able to get a straight answer to “what is ‘quantum’?”! It seems like no two physicists you ask seem to give the same answer and some don’t seem to be able to give a specific answer at all. Maybe this is just because the physicists I have asked have better things to do than memorize a dictionary definition to give people off the top of their head; but sometimes I worry that this is because most physicists actually do not themselves quite realize what “quantum” means! It seems like “quantum” is really a big group of related mathematical and scientific tools (probability, complex probability, unitary system evolution, the trick with adding differentiation operators in the hamiltonian, quantum ‘leaps’, discreteness) each of which has to do with quantum-ness but is not itself quantum-ness. I fear that a lot of physicists just call something “quantum” if it resembles one or more of these tools, but haven’t sat down and wondered, okay, what does quantum mean?
This said, Scott Aaronson offers a very good explanation of what “quantum” means in his “quantum computing since democritus” course notes. His definition of “quantum” is kind of specialized to the study of quantum computers, so maybe not all physicists would agree with it exactly. But I think it’s very close to what most physicists think “quantum” means, and anyway, I care most about quantum computers anyway
Scott kind of describes (as far as I understood his course notes) “quantum” as anything that fits the two conditions:
1. Generalized probability. Quantum processes are probabilistic, but the probabilities do not have to be real numbers as they are in traditional statistics– negative probabilities or imaginary probabilities make sense.
2. Linearity. The only operations you can perform in a quantum context are linear transformations.
These two requirements don’t tell us much from the perspective of how physical things behave, but in terms of mathematical models they say a lot.
You might actually want to read those course notes, by the way, I link part 9 but he talks about a lot of stuff besides just computer science especially at the beginning….