Can color be determined from black and white negatives?

I know that color can’t be conclusively determined from a printed/developed black and white photograph.

What I’m wondering is, is the same true for black and white negatives?

No.
That information is irretrievably lost.
If you had taken three different images of the same scene, through three different filters (RGB), than the color could be recovered - which is how Technicolor works. But a single unfiltered B&W exposure looses all the color information.
If you knew something about the scene to begin with (like, if you photographed a color chart along with the scene itself), you could make some educated guesses, and might be able to reconstruct a passable color image.

There’s been some good work done lately with AI colorizing of B&W images. But while it can do things like faces, vegetation, and the sky well enough, it doesn’t know what color things like clothing or vehicles should be.

Actually, you don’t even need that much information. Read up on the Land Effect, discovered by Edwin Land, founder of POlaroid. He came up with his retinex theory of color to explain it. If all you know is the Tristimulus Theory of color, this is guaranteed to blow your mind.

Land found that if you shot two photographs on black and white film, but looking through two different colored filters, then developed them and projected them atop each other through the same filters, it produced an image that appeared to have the full original colors. It even worked if one of the images was black and white, which comes close to what the OP is looking for.

http://www.greatreality.com/Color2Color.htm

This page tries to redo the experiment and it sort-of works, but I’ve seen much better (Land did a LOT of experiments with this, and published several journal articles. His results definitely look better than this page.):
http://www.aw3rd.us/scief/colorviz.htm

Another situation where it could be done would be if the original scene were constructed from a limited color palette. Like, if it were a paint-by-numbers picture, with the entire picture composed of eight different colors. Each color would probably (though not necessarily) be a slightly different shade of gray.

But of course, the real world isn’t like that. You’ve got red, but also light red and dark red and darkish-red-that-isn’t-quite-as-dark-as-dark-red, and so on.

Another situation where it could be done would be if the original scene were constructed from a limited color palette. Like, if it were a paint-by-numbers picture, with the entire picture composed of eight different colors. Each color would probably (though not necessarily) be a slightly different shade of gray.

But of course, the real world isn’t like that. You’ve got red, but also light red and dark red and darkish-red-that-isn’t-quite-as-dark-as-dark-red, and so on.

xx

To be clear, I’m not asking does a process to restore the color currently exist.

I guess what I’m trying to ask is…

When light strikes a black and white negative, are we certain that nothing is captured in the negative, even at the molecular level, about the wavelength of the light that hit it?

Silver halides, the light sensitive compounds in traditional photographic film, are only sensitive to the blue end of the visible spectrum. Orthochromatic and panchromatic photographic emulsions contain dye sensitizers to extend their sensitivity to longer wavelengths, giving a more natural tonal response. All modern general-purpose films are panchromatic. I don’t know the chemistry of these dye sensitizers but I would suspect that any trace colour information would be contained in alterations to these dyes and is certain to be lost in the development process. There is no colour information in the silver crystals which make up the image of a processed negative.

Which is why on most old B&W photos the sky is blown out to white, and bowls of fruit look rotten.

This is also how early color still photos were captured. The US Library of Congress has a remarkable collection of photos from early 20th-century Russia here; click the category links right above the article (architecture, ethnic diversity, etc.), then you can click on the small shots to blow them up to a larger version. In the technique that Prokudin-Gorskii used, the three color shots were taken in sequence; in photos where the subjects moved between each color shot (e.g. rising smoke or restless infants), you can see color “ghosts” where the three color images didn’t line up properly because of the movement.

I build models of military aircraft and vehicles, and there is a lot of interpretation of B&W photos (by people with far more expertise than me) to determine what colours and markings are correct, which builds on the basis of known standard paint colours to try and identify non-standard colours.

For example, the colour that an ace painted the nose of his aircraft can be worked out by comparison to the tones of the national insignia or the adjacent camouflage colours in the same photo or others taken at the same time, along with the known tonal differences of the type of film used. The result can be a very good guess as to the real colour (or the beginning of a protracted online argument).

I can imagine a situation where it might work. Suppose that you had three pigments: One (like silver halide) that turned black on exposure to blue light, one that turned black on exposure to green light, and one that turned black on exposure to red light. Mix them together in the proper proportions and coat a sheet of film with them, and that film would produce a good black-and-white image of a scene (without blowing out skies or making fruit look rotten). To the eye, that negative would just look black and white, but one could reconstruct color information by measuring which of the three pigments it was that turned black.

Except I don’t think that’s actually the way any black-and-white film works.

I can imagine a situation where it might work. Suppose that you had three pigments: One (like silver halide) that turned black on exposure to blue light, one that turned black on exposure to green light, and one that turned black on exposure to red light. Mix them together in the proper proportions and coat a sheet of film with them, and that film would produce a good black-and-white image of a scene (without blowing out skies or making fruit look rotten). To the eye, that negative would just look black and white, but one could reconstruct color information by measuring which of the three pigments it was that turned black.

Except I don’t think that’s actually the way any black-and-white film works.

If you develop Kodacolor as if it was a Black and White film, that’s what you would get.

My first “scanner” was a black and white security camera mounted above a table and connected to an Amiga 500. You would position what you wanted scanned under it, rotate a red, blue, or green filter in front of the camera, and click on the corresponding button on the capture program, which then combined the three images. My first flatbed scanner was a similar system–it had a monochrome scan head and made three passes, each with a different colored light.

Any thread where the Land Effect gets cited is a good thread.

Kind of similar, back in the day, the BBC used to syndicate TV programs internationally by converting color video to 16mm BW film by filming a TV screen and then sometimes wipe the original video tape. The way color TV works is by encoding the luminance (BW) signal in the primary bandwidth and the color signal in a subcarrier frequency. This was so that color TV would be backward compatible with BW TVs. This subcarrier would leak into the BW signal and show up as dot crawl, which was recorded on the BW film. As the dot crawl was unique to each color, it was discovered that the original color could be recovered and replaced. Several shows that only existed as BW copies got their color back using this technique.

OP, you might consider this a “sort of”:

If two people take black and white photos of the same scene but they’re using different films that have different response curves, you could reconstruct some actual color. I’m thinking in particular of the “ortho” films of earlier days and the “pan” films that largely replace them. The deal was that earlier ortho films needed more powerful photons to activate them, meaning, they responded to the shorter wavelength visible light only – the blues. Later films were created that responded to all visible photons, and were called “panchromatic” or “pan” for short. The sky might be similarly exposed (meaning similarly dark on a negative) for both kinds of films, but if a shirt was bright on the ortho and dark on the pan, it must be blue or maybe green, whereas if it was bright on both, it must be red or maybe yellow.

I’m still banking on multiple films in multiple cameras, here, but they are all BW films and it has to do with something distinctive about color being captured at the molecular level.

I bet you don’t get anything closer than this admittedly stretching-it submission, though.

If you know the light sources/conditions of the photograph and had access to the original film type wouldn’t it be possible to map each tone to a colour?