L3, I suspect you asked that question when you were very tired or had been working too long. Maybe it is time to take a nap. You are no longer thinking clearly. If you don't first take a nap, you might find yourself asleep before you finish reading my reply.
Look, it works like this, more or less.
Anders Celsius, as revised by Linnaeus, set up a temperature measurement scale starting at the freezing point of water and measuring the boiling point as 100 degrees. OK?
Well, it should be OK, because it is a nice practical scale for everyday use and for most laboratory purposes as well. By way of analogy, suppose I found that London measured ten miles across, I might decide to take the mile for my unit of measurement, and put up milestones on major roads saying "ten miles from London", "thirty miles from London" and so on. Fine. But by the time I get to the Scottish Highlands, people might not be urgently interested in the distance to London. The sign that says 500 miles to London might be merely an irritant where people would rather know that it is fifty miles to Glasgow, or something.
And as for travelling in the neighbourhood of New York!
The miles would still do fine, but the point one measures from would be an inconvenience. This would remain true, even though a tank of petrol would still take me fifty miles or whatever the distance might be, and I still would drive the same number of miles at the same speed in the same time.
Still OK?
Then I get a really smart idea. Why not set up a different set of milestones saying how far to New York or Edinburgh or the whichever suited my local needs?
Ah! But then I could not mark them in miles, could I? After all, I used miles for distances from London, right? How could I use miles from New York? Would it work?
On due consideration I conclude that maybe it would work. I would not have to buy a new tape measure or a new milometer for my car, all I would need would be to remember which city or landmark I was measuring from. The unit of distance would remain usable, and it would be a lot more convenient than having one unit of measure of distance near London, say miles, another near New York, say millimetres, another one, say rods, near Ljubljana and so on.
Right?
Now, for various purposes in physics we wish to work with the concept of "absolute zero". For mathematical reasons it happens to be much more convenient to regard absolute zero as mathematically equal to zero, just as we might regard the distance to London as being zero when we already are there, rather than in Ljubljana. (Never mind the question of just where London might be; there are quantum reasons not to be too fussy about just where absolute zero is too!) Also, absolute zero is inconveniently far from zero Celsius as well, but the unit of temperature that we use in Celsius is still quite convenient, so why not choose a new scale in which we still use units of the same size, but with zero at absolute zero, instead of at the temperature of melting ice.
No reason at all!
But which errr... theory is right then?
There is no argument about theory here. Both systems of measurement use exactly the same theorical bases. It is just that one of them measures from one point, and one measures from another. Does the "100 miles to London" sign stop being right just because there is a sign next to it saying "500 miles to Grottyville"? And if I go 20 miles Londonwards, will I be surprised to find two signs saying "80 miles to London" and "520 miles to Grottyville"? Everything is as one would expect, just measuring from what we call different points of reference.
Is it beginning to make a bit more sense after a nap?
Lets take your example. 0C.
Well say we are where we see tow signs. One says 0 miles from London and another says 273 miles from York or some such place. We walk on for twenty miles farther from York....
We find two new signs. Amazing! One says 20 miles from London. The otehr says 293 miles from York.
Now try doing the temperature calculation on the same basis.
