Dave Siemens wrote:
"On the infinite number of models, we go back to the turn of the past century. ... A country engineer, Koenigs, presented a proof that any pattern of motion could be arrived at by an infinite number of different devices."
On further thought I'm wondering if this isn't really something like the inverse of what we need to make the point. That is, Koenigs proved that an infinitude of models could produce a particular pattern of motion. But such a model is not analogous to a scientific theory. A scientific theory goes the other way: In science a single model accounts (at least in principle) for an infinitude of motions. I'd say it's vastly harder to find even a small number--much less an infinitude--of models that can adequately do this than to find an infinitude of ways to generate a specific pattern. So I'm not convinced.
Don
----- Original Message -----
From: D. F. Siemens, Jr.<mailto:dfsiemensjr@juno.com>
To: dfwinterstein@msn.com<mailto:dfwinterstein@msn.com>
Cc: asa@calvin.edu<mailto:asa@calvin.edu>
Sent: Sunday, May 01, 2005 4:18 PM
Subject: Re: definition of science
On the infinite number of models, we go back to the turn of the past century. French engineers were working on a project of designing devices to produce various motions, with the aim of having a specific plan to introduce to produce whatever motion their larger device needed. They hoped for a simple menu to solve engineering problems. A country engineer, Koenigs, presented a proof that any pattern of motion could be arrived at by an infinite number of different devices. I ran the paper down in /Comptes rendues/ some decades back, but do not remember the issue. Poincare expanded the proof in one of his books to cover any set of data that fall under the least action principle. Logico-mathematical models are more flexible than mechanical devices, which finishes the proof.
As a practical matter for supercession, theorists have to produce theories that make different predictions. However, such replacement is not necessary. It may be enough to produce a more convenient version. I have noted that some recent proofs, Fermat's last theorem and the 4-color map problem, are appallingly long. A new approach leading to an elegant proof would not prove something new, but would be hailed. From a different angle, experimentalists won't know what to look for unless a theorist has produced a tentative map. Theorists can't know that they've produced something relevant until experimentalists can test it. Denigrating the alternate approach is quite silly, but very human.
As for QM, the earliest approach to quanta involved particles--Planck, Einstein, Bohr, Sommerfeld. De Broglie devised a wave theory later, I think about 1923. I believe Schrodinger showed the equivalence of the two approaches a few years later. Since the historical development of theory is not all that important to practitioners, but solutions and applications are, you got the amalgamated approach.
When I need to check out with my groceries, the several clerks are equivalent. Each one is equally skilled in scanning the items and taking my money. They are, however, not identical. Leibnitz held the theory of the identity of indiscernibles. But it does not seem to be a necessary principle. I may not, in mathematics, be able to distinguish between 3+1 and 2+2, but there is a difference between a fried egg and a 3-egg omelet and 2 2-egg omelets. Euclid's "Things equal to the same thing are equal to each other" holds in geometry, but not necessarily for comestibles.
Dave
On Sun, 1 May 2005 06:23:15 -0700 "Don Winterstein" <dfwinterstein@msn.com<mailto:dfwinterstein@msn.com>> writes:
DFS: I have noted that there are an infinite number of theories/models. To expand, some of them will be exact matches (QM alternates).
My thesis chairman used to say, "Theorists can explain anything in a variety of ways. It's up to experimenters to say who's right." He was an experimenter (as I also became). To be relevant, theorists need to be able to make predictions that distinguish their theories from others. I suspect theories that are exact matches in that all their predictions are identical to those of another theory are in fact identical theories even though expressed differently. In other words, it should be possible to show how two theories that predict the same things are alternate ways of saying the same thing.
You say, "Philosophers insist that equivalence is not identity." I'd need proof. In classical mechanics we have various formulations such as those by D'Alembert, Lagrange and Hamilton, but I've never heard anyone call these different theories; and in fact I'm sure it's possible to show they're all different ways of saying the same thing.
Going back to your comments on string theory: No doubt the math is very complicated (but I haven't looked at it, so I can't speak from experience). Greater complications are to be expected when the scope of the theory is greater. General Relativity is more complicated than Newton's law of gravitation, but it explains more. String theory tries to explain a lot more. But complicated as these theories are, in terms of Einstein's previously cited essay they would be simple in Einstein's usage.
(But string theory at this stage is still a branch of math. Not until it makes testable predictions can it be considered physics. Science requires a hard connection between the idea and the real world.)
QM as I understand it, however, would not be simple in that sense. Robert Dicke (with James Wittke) was the author of my undergraduate QM text, which lists seven "postulates of quantum mechanics." These are essentially assumptions that relate the math to the physics. To give the flavor, his postulate 1: "It is assumed for a system consisting of a particle moving in a conservative field of force (produced by an external potential) that there is an associated wave function, that this wave function determines everything that can be known about the system, and that it is a single-valued function of the coordinates of the particle and of the time." So QM is not derived from basic principles but simply assumes a lot of stuff at the outset. This is not significantly different (I think) from the "piecemeal curve fitting" I mentioned earlier (and possibly gave Einstein fits).
DFS: My impression from outside is that the "particle" theory and the "wave" theory give identical results for all computations.
DFW: OK, I have to say I don't know where you're coming from here. I've had several graduate-level courses in QM but have never heard of this.
Don
----- Original Message -----
From: D. F. Siemens, Jr.<mailto:dfsiemensjr@juno.com>
To: dfwinterstein@msn.com<mailto:dfwinterstein@msn.com>
Cc: asa@calvin.edu<mailto:asa@calvin.edu>
Sent: Saturday, April 30, 2005 12:34 PM
Subject: Re: definition of science
As I recall, I have noted that there are an infinite number of theories/models. To expand, some of them will be exact matches (QM alternates). Others will match up to a point (Whitehead and Einstein). The impression I got is that some of the theories mentioned in the SciAm article belonged in each of these categories--Dicke's in the latter.
Call it what you will, a particle is not a wave, and neither are wavicles. A hole can be punched with a bullet or with a laser. But they are not the same. My impression from outside is that the "particle" theory and the "wave" theory give identical results for all computations. However, some computations that are devilishly difficult in the one are relatively simple in the other, but there are two theories for the same results. The pair, because equivalent, are used as one by scientists. Philosophers insist that equivalence is not identity. Scientists round off and substitute linear approximations for nonlinear equations. Mathematicians say it's wrong. This depends partly on the fact that scientists have to measure, necessarily with limited accuracy, while mathematicians submit to a perfect ideal. After measuring with limited accuracy, scientists idealize their results. Empiricism has its requirements, as do formalisms.
Dave
On Sat, 30 Apr 2005 01:41:02 -0700 "Don Winterstein" <dfwinterstein@msn.com<mailto:dfwinterstein@msn.com>> writes:
DFS: Are you suggesting that only one mathematical model fits? ..
DFW: Einstein believed, as do I, that one is superior to all others. Only experiment can decide. (Although, if you're Einstein, you can apparently decide also on the basis of which theory has the simpler math. : ) ) RH Dicke also presented an alternative to Einstein's General Relativity, as I recall. I think he wrote the SciAm article. But none of the other versions have earned widespread acceptance.
DFS: Quanta may be approached either as particles or as waves, equivalent theories. ....
DFW: There are different representations of the QM formalism (e.g., Dirac's bra & ket notation), but it's the same theory. Lots of people would like a more "reasonable" theory, and I've heard of attempts, but no one has yet improved on QM as we know it. As for waves and particles, it's not that we're invoking different but equivalent theories but that we're doing different kinds of experiments. Particles are also waves; but when measuring them, some experiments detect particle-like behavior and others wave-like behavior. It's called complementarity. Or don't I understand what you're saying?
Don
----- Original Message -----
From: D. F. Siemens, Jr.<mailto:dfsiemensjr@juno.com>
To: dfwinterstein@msn.com<mailto:dfwinterstein@msn.com>
Cc: asa@calvin.edu<mailto:asa@calvin.edu>
Sent: Thursday, April 28, 2005 11:02 AM
Subject: Re: definition of science
Are you suggesting that only one mathematical model fits? After Einstein presented his work, Whitehead came up with a different version. Not liking Riemannian geometry, his was based on Euclidean. Eddington proved that the two were equivalent on the four matters then recognized as relevant. Later work disproved Whitehead's version of relativity because of other matters. I recall an article in /Scientific American/ that presented additional relativity theories, though it did not discuss the calculus underlying them. Some apparently were equivalent to Einstein's theories, while others were designed to be slightly different.
Quanta may be approached either as particles or as waves, equivalent theories. I understand that two approaches to string theory were demonstrated equivalent. In other words, the fit is multiple. Beyond that, are four dimensions simple? What about 10 or 11? Is seeing a matching pattern simple? Once seen, it's "obvious," of course. Then why does it take brilliant people so long to see it? How many have an /annus mirabilis/?
Dave
Received on Wed May 4 03:49:45 2005
This archive was generated by hypermail 2.1.8 : Wed May 04 2005 - 03:49:45 EDT