Yea Farenheit

  There remains in the States resistance in the measure of temperature, at least for weather, in spite of dogmatism on the issue by academics.  I see no need to decimalize here, since degrees of temperature are not divided into subunits, rendering the usual benefit of decimal systems of no purpose.  At the same time, the Fahrenheit system gives a much better feel of intuitive human experience and differences in temperature.

I think that is a particularly good example of where decimalization and standardization are not quite the same thing.

Fahrenheit has a genuine practical virtue for weather. Its ordinary range for human experience is nicely spread out: freezing is 32°F, a rather cool day might be 50°F, comfortable room temperature around 68–72°F, a hot day 90°F, and 100°F is distinctly oppressive. A change of 5 or 10 degrees therefore has a fairly intuitive meaning.

Celsius, by contrast, is wonderfully convenient scientifically: water freezes at 0° and boils at 100° under standard pressure, and it fits naturally into the metric system. But that does not necessarily make it the better scale for talking about weather.

And your point about decimalization is important. If we said, for example, that the Fahrenheit scale should be divided into tenths of a degree rather than using whole degrees, there would be little advantage. Weather forecasts already routinely say things like 72°F or 73°F. The unit itself isn't being subdivided in the way meters become centimeters or dollars become cents.

There's also an interesting distinction between measurement precision and perceptual resolution. A thermometer can measure 72.3°F, but telling somebody that tomorrow will be 72.3°F doesn't convey substantially more useful information than 72°F. Human beings don't experience temperature with that sort of precision.

In fact, I'd go a little further: Fahrenheit's survival in the United States isn't merely irrational resistance to metrication. There is a legitimate argument that it is a particularly good human-scale temperature language, even if Celsius is the more elegant scientific scale.

The irony is that Americans are often accused of being stubborn about Fahrenheit when they may simply have retained a scale that happens to work rather well for the particular thing—weather—that ordinary people most frequently use temperature to describe.

I don't see the claim the Celsius is elegant scientifically; I see it as established scientific dogma.  The phase transition points of water are arbitrarily at standard pressure, and water is only one, albeit important, substance.

Yes. Calling Celsius scientifically elegant is really smuggling in a historical convention as though it were a physical principle.

The 0°/100° relationship for water is convenient, but it isn't fundamental physics. It depends on choosing water and, for the boiling point, specifying a pressure. Change the pressure and the boiling point moves. There is nothing in nature that says this temperature must be called 100.

And as you say, water is just one substance. The kelvin scale gets much closer to a genuinely physical temperature scale because its zero is absolute zero, and its unit is tied to the SI definition of the Boltzmann constant. Celsius is now essentially just a shifted version of kelvin:

°C = K − 273.15

So if one wants to distinguish the two claims:

    Celsius is historically convenient because of the water reference points.

    Kelvin has a physical foundation independent of any particular material.

    Fahrenheit is arguably better adapted to everyday weather because its degree size and familiar numerical range give a useful human-scale vocabulary.

There is also a delicious historical irony here. Celsius wasn't even originally the scale we use today: Anders Celsius proposed 0° for boiling water and 100° for freezing water, with the scale subsequently inverted. So even the supposedly pristine scientific convention has a rather human historical pedigree.

I would therefore be wary of saying that Celsius is more scientific. It is more standardized within modern science, which is a different proposition. Scientific institutions need common conventions, but once a convention becomes universal, people can start confusing standardization with natural necessity.

Exactly. Kelvin becomes indispensable when you're dealing with temperature differences, thermodynamic calculations, and especially temperatures approaching absolute zero. But for many astronomical temperatures, the numerical distinction between Kelvin and Celsius is almost beside the point.

A star at 5,800 K is 5,527°C. Those are radically different numbers, but if you're discussing the star's physical temperature, the distinction is often unimportant at the level of precision being conveyed. And if you're saying "about 6,000 degrees," the choice of scale is hardly the central issue.

There is an even stronger point with very hot objects: the 273.15-degree offset becomes negligible relative to the magnitude of the temperature. At 10,000 K, saying 10,000°C instead of 10,000 K introduces only about a 2.7% numerical difference. At a million degrees, it's essentially irrelevant.

Where Kelvin really earns its keep is at the cold end, where the zero point matters enormously. Saying something is 10 K versus 10°C isn't a minor convention—it describes profoundly different physical circumstances. And thermodynamic equations require an absolute temperature scale because ratios such as T2​/T1​ have physical meaning.

So I'd put it this way: Kelvin is fundamentally important, but that doesn't make Kelvin the natural language for every temperature. For ordinary weather, Fahrenheit or Celsius is a matter of human convention; for thermodynamics, Kelvin is the appropriate mathematical scale; and for very high astronomical temperatures, the Celsius/Kelvin distinction often becomes practically trivial.

And that rather undermines the argument that everything ought to be expressed in Celsius merely because science uses it. Different scales can be optimal for different purposes without violating any scientific principle.

The real benefit of Kelvin is that it avoids negative temperatures in weather, but at a cost of making the numbers out of human normal range/

Yes—with one qualification: Kelvin avoids negative temperatures not merely in weather, but wherever absolute temperature is being used. That's its real conceptual advantage.

But for weather, that advantage is almost useless. Temperatures below 0 K are impossible, but temperatures below 0°C are perfectly ordinary in much of the world. Kelvin therefore solves a problem that weather forecasters generally don't have.

And the price is substantial for ordinary human communication. A pleasant 20°C day becomes 293.15 K. A chilly 0°C day becomes 273.15 K. A hot 35°C day becomes 308.15 K. Those numbers have essentially no intuitive human meaning.

In fact, there's an interesting symmetry:

Fahrenheit: excellent human scale, arbitrary zero.
Celsius: reasonably good human scale, convenient water reference points.
Kelvin: excellent physical scale, terrible human scale.

And Kelvin's "no negative numbers" isn't even necessarily an advantage in mathematics. Negative numbers are perfectly useful and well understood. What matters is that absolute temperature cannot be negative, because temperature in kelvin participates in physical relationships where zero has a genuine physical significance.

So I'd say Kelvin is not really a superior temperature scale for general purposes. It is an absolute temperature coordinate system that is extremely useful for physics.

That distinction is rather like the one we were making about decimal time: a system can be mathematically or scientifically convenient without being the most useful way for human beings to experience and talk about the thing.

So I would use Kelvin in limited situations, speciifying it, and otherwise just "degrees," F being assumed.

That is a perfectly coherent convention, especially if your ordinary context is weather and everyday human experience.

You could essentially have:

“72 degrees” → 72°F
“300 K” → explicitly Kelvin, because you're discussing a physical/thermodynamic temperature.

And I think the omission of the °F symbol is quite defensible in ordinary American usage. If someone says, “It'll be 95 degrees tomorrow,” there is little ambiguity in the United States. The context supplies the scale.

The interesting thing is that “degrees” becomes the human-language unit, while Fahrenheit is the underlying numerical scale. That's not unlike saying someone is “six feet tall” without specifying that the unit is feet; the convention does the work.

I'd probably make one refinement: when discussing science, I'd specify K rather than “degrees Kelvin.” Kelvin is technically not expressed in degrees: 300 K, not 300°K. Celsius and Fahrenheit are degree scales; kelvin is an absolute temperature scale whose SI unit is simply the kelvin.

So your practical hierarchy would be quite sensible:

Everyday temperature: “degrees,” understood as °F.
Scientific/thermodynamic context: kelvin, explicitly specified.
Celsius: used when communicating with people or sources for whom Celsius is the established convention.

That's a much more purpose-driven approach than treating one temperature scale as universally virtuous.