Michael Edmiston raises a point of terminology. As I did my PhD in
experimental solid state (now condensed matter) physics I also used
the terms "ohmic" and "nonohmic" in relation to contacts. In this
discussion the term is used with respect to a device, and it means
(transparently, I thought) that the device does not obey Ohm's law.
I will point out that temperature is not the problem here. One can,
if one is perverse, immerse the bulb in a constant temperature bath
and it will still disobey this "law" of Nature
Michael implicitly assumes that Ohm's law is a statement about the
response of matter, that it is what physicists mean when they write
J = sigma E
As we know well, this is not a law of Nature either; it is merely
the first term which is dominant in materials we describe as linear.
In pure crystalline materials this strictly local law must again be
modified, since the current density at any point depends also on
the electric field in the surroundings. I could go on (and I did in
my qualifying examination) but I'm sure you get the idea.
I am dealing with a device-based Ohm's law (the one in question in
this discussion) which is usually written as
V = I R
or some near variant thereof. (I'm sure many of us remember the TPT
cover with the blackboard showing the three laws of electricity.)
I think that is the extent of Michael's disagreement with what I
have said. If it is not, I await correction. I will eschew the use
of the terms "ohmic" and "nonohmic" in this context if they are at
all confusing to anyone reading this.