A little math quiz

Um, perhaps the confusion is about the potential energy per unit distance, vs the potential energy in a single winding.
I think what I said in #75 and #76 is that you integrate over distance, but since the windings are farther apart at one end, the potential per winding is greater at that end--the weighted end. Sorry if it looked like I was saying the potential per unit of distance along the spring varies, because that has to be constant if k is constant everywhere along the spring.

It's why I said you could integrate over units of tension because these are equivalent to units of distance (in a "spring" rest frame)
 
Returning to mathematics......

Yes it's ironic that the OP has hijacked their own thread :)

I assert that then $$\mathbb{F} \otimes V_n \simeq \mathbb{F}^n$$

I would like to make a notational quibble (two actually) I regard as important.

I'm just trying to make sense of tensor products so nothing I say here is authoritative, I'm just stumbling around in a dark room as Wiles put it.

What I think you mean to say above is this:

"I assert that then $$\mathbb{F} \otimes_\mathbb F V_n \simeq \mathbb{F}^n$$ as vector spaces." Or do you mean $$\otimes_\mathbb V$$? I'm not actually sure which you mean here.

There are two separate things going on here. First, the statement that this is a vector space isomorphism as opposed to, say, an algebra homomorphism.

This is key because in the subject of this thread, $$\mathbb C$$ is isomorphic to $$\mathbb R^2$$ as $$\mathbb R$$-vector spaces; but not as algebras, since $$\mathbb R^2$$ has no algebra structure at all. That is in fact the essential difference between $$\mathbb C$$ and $$R^2$$.

An algebra is a vector space where there is a multiplication defined on the vectors that's compatible with the underlying scalar multiplication. The classic case is $$\mathbb C$$ regarded as a vector space over $$\mathbb R$$. It's true that we have a vector space, but of course something much stronger is true as well. We have a multiplication defined on $$\mathbb C$$ that makes $$\mathbb C$$ into a ring (in this case a field); and this multiplication is compatible with scalar multiplication by elements of $$\mathbb R$$.

For example $$\alpha (z + w) = \alpha z + \alpha w$$, and this is true whether $$\alpha $$ is real or complex; and whether we regard this as scalar multiplication by a real; or multiplication by a complex number that happens to be real.

It's this extra algebra structure that makes the complex numbers different from the reals. https://en.wikipedia.org/wiki/Algebra_over_a_field

So first, it's important to say whether your isomorphism is of vector spaces, algebras, or something else (rings, modules ...).

The second missing piece of notation is a subscript on the tensor product sign that identifies the underlying field (or ring). This is important because we are interested in extension of scalars, which is when we take an $$\mathbb R$$-vector space and use a tensor product to make it into a $$\mathbb C$$-vector space. In this case saying $$X \otimes_\mathbb R Y$$ is very different than $$X \otimes_\mathbb C Y$$

Along these lines I wanted to mention that in the very first post of this thread, we have:

It's well known that the complex plane $$ \mathbb C $$ is isomorphic to $$ \mathbb R^2 $$

In fact this is not well known to me! Are these two objects isomorphic as Abelian groups? Yes. As rings? No. $$\mathbb R $$-vector spaces? Yes. $$\mathbb R $$-algebras? No. $$\mathbb C$$-vector spaces or algebras? No.

It is important to be clear about the type of isomorphism (between what type of objects) and if there's a base field involved, what is it?
 
Last edited:
someguy1 said:
Yes it's ironic that the OP has hijacked their own thread :)
I don't think that's the case. The two experiments exemplify differences between linear kinds of physical space (springs at equilibrium), and nonlinear kinds (particles interfering).

In the weighted springs frame, you have the relation mg = kx, or mg - kx = 0, an equation in two variables, i.e. an ordered pair of scalar quantities (m,x). These are obviously in $$ \mathbb R^2 $$. But particles are in $$ \mathbb C^2 $$ ('tHooft tell us electrons need two complex numbers to describe spin); the Schrodinger eqn in one dimension is in $$ \mathbb C $$. The last is enough dimensions to define tunnelling through a potential barrier.
 
I think the biggest difference between $$ \mathbb R^2 $$ and $$ \mathbb C $$ is differentiability. Related (I'm not sure about the details, it's a bit advanced) to the Cauchy-Riemann condition, which is about continuity and the existence of a limit.
 
I think the biggest difference between $$ \mathbb R^2 $$ and $$ \mathbb C $$ is differentiability. Related (I'm not sure about the details, it's a bit advanced) to the Cauchy-Riemann condition, which is about continuity and the existence of a limit.

That's interesting. I've seen Cauchy-Riemann derived by letting the variable approach the origin along the real and imaginary axes and equating the two results.

In the proof on Wiki, a key part of the argument is to use the fact that you can multiply or divide a complex number by $$i$$. So this does seem to relate to the multiplicative structure on $$\mathbb C$$. Once you can multiply by $$i$$ you get the field structure, which $$\mathbb R^2$$ lacks. [Field as in abstract algebra, and NOT field as in physics, which is totally unrelated].

The article also talks about the physical basis of the Cauchy-Riemann equations. They're related to fluid flows on the plane and the Laplace equations. It has something to do with the fluid being incompressible. I do wish I understood more about the physical basis of math.
 
Last edited:
Damn. I just realised I've got it wrong about the distance between windings. This distance must be the same over the whole spring at equilibrium--I've been thinking about what happens when the spring accelerates.

But, I recall that hanging a spring vertically so it extends under its own weight, does sometimes give different spacing between the windings.
It depends somewhat on the length (its more apparent with a longer spring, hence a smaller k), but what happens is the reverse of what I've been saying: the windings are separated near the top of the spring, but stay close together near the bottom. The upper part of the spring is supporting the weight of the lower part, type of thing as the spring "weighs itself".

It would be interesting to relate the differences in spacing to a sum over the mass of each winding (notwithstanding the difficulties with measurement accuracy).
 
Maybe you're thinking, the spacing you see in a spring hanging vertically under its own weight, is because the spring isn't all that homogenous, there is a bit of change in tension here and there. That means you should see something different if you swap the ends of the spring around, but you don't.

So, you want to sum over n windings. At the top of the spring, the first winding is supporting the mass of n-1 windings, the first two windings support the mass of n-2 windings and so on down. Therefore the first winding is extended the farthest because it supports most of the mass of the spring. If you know the k of the spring you should be able to calculate the k of a smaller piece of the same spring, say a few windings long. So you can calculate the mass in n - 1, n - 2, ... windings, so you can use the spring to weigh the spring.
 
Um, perhaps the confusion is about the potential energy per unit distance, vs the potential energy in a single winding.
I think what I said in #75 and #76 is that you integrate over distance, but since the windings are farther apart at one end, the potential per winding is greater at that end--the weighted end. Sorry if it looked like I was saying the potential per unit of distance along the spring varies, because that has to be constant if k is constant everywhere along the spring.

It's why I said you could integrate over units of tension because these are equivalent to units of distance (in a "spring" rest frame)

I have not checked but, in equilibrium the elongation per unit length should decrease as we move up from force end to fixed end...Could this be because of change in k (as we move towards fixed end, the smaller equivalent length will have proportionally higher k, thus lesser dx).
 
The God said:
I have not checked but, in equilibrium the elongation per unit length should decrease as we move up from force end to fixed end...
No, what I was saying isn't true, in fact the elongation between windings is the same at equilibrium--each winding contributes an equal amount of tension to the overall linear tension constant k.

It's only true that you get different spacing between windings when the spring is hanging vertically under its own weight.
What you can do in terms of measurement with a spring is: measure the rest length and changes in length under a constant force, count the number of windings, although this may not be an integral number, and you can weigh the spring.

Different weights attached to the end of the spring then each add a constant force, F = mg, if the weight is hanging freely, so you have a way to calculate the k of the spring via changes in its total length. Then if you divide the rest length of the spring by the number of windings, you have the contribution each winding makes to the rest length, then if you extend the spring so you have mg = kx, you can divide x by the number of windings to get the contribution each winding makes to x.
You also have that mgx is the work done on the weight as it gets lowered to its equilibrium position.

So I've thought that the original experiment I did must have used a set of springs that were all the same spring with different rest lengths. 'Normalising' the k of different springs (with, say, different mass or different winding radius) is a harder problem. Otherwise, any plot of mg vs kx will have the same y intercept for a set of springs which are materially identical except for their rest length.

So what is this constant, what does measuring springs under applied forces tell us about it?
 
No, what I was saying isn't true, in fact the elongation between windings is the same at equilibrium--each winding contributes an equal amount of tension to the overall linear tension constant k.

Put the spring on Horizontal Table top and apply a constant force in Horizontal direction, that will take care of gravity part....

Now divide the spring in 'n' parts, the constant of each part is 'nk'.....where k is the spring constant of the original spring..will you still get the same elongation in each part even when it is in equilibrium with force F applied on one end and another end is fixed...
 
The God said:
Now divide the spring in 'n' parts, the constant of each part is 'nk'.....where k is the spring constant of the original spring..
If you add two springs with the same k together in series , then k' is given by: 1/k' = 2/k, i.e. k/k' = 2.
Which is exactly what you do if you add the k of two windings together. For n windings it's k/k' = n, where k is the spring constant for one winding, k' is the overall constant.

And, you can use what's called the spring compliance, c = 1/k, so you also have c'/c = n. And you should remember that Hooke's Law applies: the spring has to be within its linear range of operation (no overstretching).

Because you have a relation, mg= kx, you can plot it in $$ \mathbb R^2 $$, because m in kilograms, and x in metres (or whatever units you choose for mass and distance) are orthogonal. Lots of things you can plot in physics are orthogonal, often the x axis is where time values are plotted vs just about anything else. Why not plot mg = kx in $$ \mathbb C $$? And what about when the spring + weight isn't in equilibrium, but oscillating?
 
.......do you agree or disagree that extension per unit length across the length will be same throughout ?
 
The God said:
do you agree or disagree that extension per unit length across the length will be same throughout ?
If the spring isn't accelerating it should be uniformly extended, but that depends on the spring being homogenous, i.e. a simple spring.
 
So a spring can be a linear system (a straight line plot of mg vs kx), or if accelerations are introduced, a nonlinear system where the endpoint of the spring oscillates.
Basically, this is the difference between lowering a spring at "constant" velocity so the weight does not accelerate (a damping term), or allowing the weight to accelerate freely, past the point of equilibrium. The mathematics required to describe the SHM is now a second order differential equation with complex solutions.

Now the system oscillates with a regular period, so time is part of the measurement. In equilibrium, there is a timeless state, although reaching equilibrium does involve doing work which must be in the time domain--lowering a weight in a gravitational field at constant velocity is a linear operation on the system too.
 
If the spring isn't accelerating it should be uniformly extended, but that depends on the spring being homogenous, i.e. a simple spring.

Say it not accelerating, it is homogenous...now if the total elongation is x, will it be x/2 for both the halves (one closer to hinge and another closer to force end)
 
The God said:
Say it not accelerating, it is homogenous...now if the total elongation is x, will it be x/2 for both the halves (one closer to hinge and another closer to force end)
Yes, the spring is linear and stretching it is a linear operation. When you have SHM, the windings in the spring aren't linearly spaced during acceleration, this is observable.

So, why can you plot the linear relation in $$ \mathbb R^2 $$, but you need complex numbers for SHM? Is it generally true that wavelike behaviour is complex?

Another quiz question: suppose you get asked to show that mass and distance in space are orthogonal, using vectors. How do you go about it? Can you use that space and time are orthogonal?
 
Last edited:
Another quiz question: suppose you get asked to show that mass and distance in space are orthogonal, using vectors. How do you go about it?
You can't - mass and distance are scalars. Orthogonality is a property of vectors in a metric space. I.e. Two vectors are orthogonal if their inner product is zero.
 
Well sure, but a distance with a direction, which is what linear spring extension amounts to, must be vectorial. I think what the question should have asked is can you make mass and distance look like the same scalar quantity, or are they always mapped to orthogonal vectors by any physical relation between them.
 
I notice someone posted something about space having two measurable properties, even when no matter is present. But measurement implies particles interacting and the presence of matter.

Is it possible to measure anything without doing work? If you do work are forces always involved and hence, a vector space?
 
Is it possible to measure anything without doing work?
Ah, of course, in entanglement measurement you can determine with 100% certainty the state of a particle, without measuring it (interacting with it in any way). Instead you measure its entangled partner, and you seem to get something for nothing (i.e. no work is done on what amounts to 1/2 the system).

But, you do measure one particle, you don't (or someone else does) measure the other. Determining the state of a particle without measuring it directly (although a measurement is necessary) seems to us to be an extra degree of freedom for quantum information.
 
Last edited:
Back
Top