what are the requirements of a scientific theory?

[swansont]

A theory is false with just one experiment the contrary.

 

It would be a contradiction in terms to say it should be possible in principle to find observations showing a theory to be untrue, and at the same time saying a theory is true. Indeed, if such were the case, the world would be chaotic.

A valid experiment can simply limit the scope of a theory or law. Ohm's law is not false even though there are devices which do not follow it. Relativity is not false even though GR cannot hold at very small scales. Even Newtonian gravity is not generally considered "false".

[Griffon]

A theory is false with just one experiment the contrary.

 

It would be a contradiction in terms to say it should be possible in principle to find observations showing a theory to be untrue, and at the same time saying a theory is true. Indeed, if such were the case, the world would be chaotic.

 

A so called supernatural explanation is a falsification.

I'm not sure what you're saying exactly.

 

I agree a theory can be shown to be false by experimental result.

 

My point is that not all explanations can be shown to be false. For such explanations there is no observation or experiment you can make that can show the explanation to be false. The theory that an undetectable goblin moves my car keys about so I cant find them is such an explanation. [if only I could catch the blighter though!]

 

Such explanations are unscientific. For an explanation to be described as scientific it must be possible, in principle, to show that it is false. The discovery that neutrinos travelled faster than the speed of light would have been sufficient to falsify relativity as it stands. That's what's meant by "falsifiable".

[Delbert]

Ohm's law is not false even though there are devices which do not follow it.

Name one?

 

Even Newtonian gravity is not generally considered "false".

Newtonian gravity is false. But the error is very small for everyday activity, and unless you want to calculate things like Mercury's orbit, it can be used for most things. But that doesn't make it true.

 

If you're saying 'generally considered' to mean it can be used for most things as above, then that's okay. But that's not how a theory is deemed to be true or false. And true or false is what I understand this subject to be about.

[ajb]

Newtonian gravity is false. But the error is very small for everyday activity, and unless you want to calculate things like Mercury's orbit, it can be used for most things. But that doesn't make it true.

 

If you're saying 'generally considered' to mean it can be used for most things as above, then that's okay. But that's not how a theory is deemed to be true or false. And true or false is what I understand this subject to be about.

Theories are not shown to be "true" or "false" as such, but rather they are shown to be "good" or "bad". A theory is "good" if it agrees with nature well; note that we need to be careful with the domain of validity, experimental errors and the level of accuracy in the predictions we will accept.

 

A theory is "bad" if it does not agree well with nature.

 

So, Newtonian gravity is a good theory, provided we do not have objects moving very fast nor very strong gravitational fields. For example, sending people to the Moon only requires Newtonian gravity rather than general relativity.

Name one?

Semiconductor diodes.

 

Many devices do not have a linear relation between voltage and current.

[swansont]

Name one?

 

 

ajb already took diode so I'll say transistor.

[Delbert]

Theories are not shown to be "true" or "false" as such, but rather they are shown to be "good" or "bad". A theory is "good" if it agrees with nature well; note that we need to be careful with the domain of validity, experimental errors and the level of accuracy in the predictions we will accept.

 

A theory is "bad" if it does not agree well with nature.

I was replying to Griffon when he referred to a theory as being untrue. So I used 'true' and 'untrue' in reply. It wouldn't be terms I would otherwise use.

 

Anyway, in reply to this 'good' or 'bad' business I'd look to Richard Feynman: If it disagrees with nature, experiment or experience then it is wrong. There's no mention of sometimes disagrees, good or bad qualities.

 

 

Semiconductor diodes.

 

Many devices do not have a linear relation between voltage and current.

The resistance of a diode will be equal to voltage across it divided by the current following through it i.e. R = E/I. If one passes a different current through said diode (forward biased) then the resistance will still be in keeping with the (new) voltage across it R= E/I.

 

As we all know, its resistance changes with different currents (it would be pretty useless as a diode if it didn't), but its resistance at any given voltage and current (any point on its characteristic curve) will be R =E/I. There is no contradiction with ohm law.

 

The same applies to all semiconductors.

 

 

ajb already took diode so I'll say transistor.

Ditto above.

[Strange]

 

ajb already took diode so I'll say transistor.

 

I'll go for incandescent light bulb.

The resistance of a diode will be equal to voltage across it divided by the current following through it i.e. R = E/I. If one passes a different current through said diode (forward biased) then the resistance will still be in keeping with the (new) voltage across it R= E/I.

 

That is a pretty useless redefinition of Ohm's law. If there is no fixed relationship between V and I then it isn't much of a law. Of course there are devices where the current is not constant with a constant voltage so they don't obey even your "weak Ohm's law".

 

(Are we getting off topic...)

[ajb]

Anyway, in reply to this 'good' or 'bad' business I'd look to Richard Feynman: If it disagrees with nature, experiment or experience then it is wrong. There's no mention of sometimes disagrees, good or bad qualities.

Basically this is right, but you have to carefully think about what it means to agree or disagree with nature. The best one can really do is see if your theory is consistent with nature and this can only ever be up to experimental errors and the accuracy of your prediction. Remember we also have the domain of validity, that is there will be a range of parameters for which one would not expect the theory to agree well with nature.

 

Taking this into account, if is unfair to say that, for example Newtonian gravity is wrong. It is better to say that general relativity is a better theory as it agrees closer with nature for a wider range of phenomena than Newtonian gravity. Also in this example, we can see Newtonian gravity as a limit of general relativity, so the two theoreis are consistent with each other.

 

The resistance of a diode will be equal to voltage across it divided by the current following through it i.e. R = E/I....

Well, if you plot V and I you don't get a straight line, so the theory is nonlinear. I would say that this means the device does not obey Ohm's law. But okay...

[Delbert]

 

I'll go for incandescent light bulb.

 

Yes, it's going off topic but I can't help asking what on earth is it about an incandescent light bulb that is contrary to OHM's law?

 

Well, if you plot V and I you don't get a straight line, so the theory is nonlinear. I would say that this means the device does not obey Ohm's law. But okay...

For each point on the graph to which I presume you refer, E, I & R are in keeping with OHM's law. That's what the curve represents: it's resistance in keeping with OHM's law at all points. It's different at different points as we all know, but at any point you choose the current and voltage will the result of its resistance at that point according to OHM's law.

 

For example, are you saying that if I pass a current (say) of 1amp through a diode (forward biased) and then measure the voltage across it to be (say) 0.01 Volts, the resistance of the diode won't be 0.01 OHMs? Because it will be 0.01 OHMs, and that's all OHM's law says.

 

Even a fixed resistor will change resistance to some degree with different applied voltages and corresponding currents - if for no other reason than because of the heat created. So, if it was the case you were correct, then OHM's law wouldn't even apply to a fixed resistor.

 

Anyway, I think this has drifted off topic.

 

[Strange]

Yes, it's going off topic but I can't help asking what on earth is it about an incandescent light bulb that is contrary to OHM's law?

 

The resistance is quite strongly temperature (and, therefore, voltage x current) dependent. You may not even measure the same resistance for a given voltage at different times.

 

 

For each point on the graph to which I presume you refer, E, I & R are in keeping with OHM's law. That's what the curve represents: it's resistance in keeping with OHM's law at all points. It's different at different points as we all know, but at any point you choose the current and voltage will the result of its resistance at that point according to OHM's law.

 

That is a wrong (and useless) definition of Ohm's law. A better definition is that on Wikipedia:

 

Ohm's law states that the current through a conductor between two points is directly proportional to the potential difference across the two points. Introducing the constant of proportionality, the resistance, one arrives at the usual mathematical equation that describes this relationship: I = V/R.

 

(emphasis added)

If R is not constant then it isn't obeying Ohm's law. All you are saying is that (for some devices) a fixed voltage produces a fixed current. But, of course, even that isn't true for all components.

[Delbert]

 

Ohm's law states that the current through a conductor between two points is directly proportional to the potential difference across the two points. Introducing the constant of proportionality, the resistance, one arrives at the usual mathematical equation that describes this relationship: I = V/R.

So what's the difference to what I said? The only difference as I see it is I used E instead of V. E being Electro Motive Force (EMF), more commonly know as volts i.e.V. And transposing the formula: R = E/I. Or using your nomenclature: R = V/I.

 

If R is not constant then it isn't obeying Ohm's law. All you are saying is that (for some devices) a fixed voltage produces a fixed current. But, of course, even that isn't true for all components.

 

Don't think it says anything of the sort. And if I've understood exactly what you're saying, then I think you will find that there isn't a single resistive component that would have a fixed and unchanging resistance with different applied voltages to satisfy your apparent understand of OHM's law. Even a 'fixed' resistor I understand will vary slightly with different applied voltages. And that's not considering such variations in resistance occasioned by heat - they dissipate energy as heat. And as such they are rated or classified in Watts as will as resistance.

All I've said is that the voltage across said diode divided by the current through it (forward biased) will equal it's resistance, R = E/I, which is OHM's law. And measuring the voltage for all reasonable currents will give the characteristic curve - it's doubtless better to apply a voltage and measure the resultant current (small) in the reverse biased direction. The resulting characteristic curve represents its resistance over the normal operating range of the component.

And returning to fixed resistors, I further understand with precision resistors you may receive a chart with a characteristic curve - not as dramatic as a diode curve, but one constructed in the same way displaying slightly different resistance with different voltages.

Anyway, it's clear that we agree to differ. So perhaps it's best we leave it at that.

Edited August 2, 2013 by Delbert

[Greg H.]

So what's the difference to what I said? The only difference as I see it is I used E instead of V. E being Electro Motive Force (EMF), more commonly know as volts i.e.V. And transposing the formula: R = E/I. Or using your nomenclature: R = V/I.

 

 

By definition, Ohm's law states a direct relationship between the voltage, the resistance, and the current. That is, there is a constant of proportionality between the current and the voltage. If the current varies, the voltage will vary by a proportional amount, and vice versa.

 

 

pecifically, Ohm's law states that the R in this relation is constant, independent of the current.^{1 (Emphasis mine)}

 

 

 

However, as others have pointed out, there are a great number of materials that do not obey this definition because there is no fixed relationship between the amount of current and the amount of voltage. The constant of proportionality varies as the current changes. So these materials do not follow Ohm's law, because the R factor in the equation varies in conjunction with the current. The equation applies, and can be used to determine the various terms, assuming you know any two of them, but Ohm's Law itself does not hold true, because the materials in question do not maintain a constant proportion.

 

Look up the term non-ohmic material.

 

1: Oliver Heaviside (1894). Electrical papers 1. Macmillan and Co. p. 283

Edited August 2, 2013 by Greg H.