[rachofsunshine]

The freezing-into-a-ball-of-ice is relevant here. A body that small can't hold on to water vapor at anything a human would consider a reasonable temperature; the average velocity of light gases at human-sane temperatures is high enough to overcome their escape velocity. See [1] for a log-log plot of what gases a body can hold onto - even Mars, which is much larger and denser than a Ceres-sized ball of water, has lost most of its water (although other factors like the solar wind are contributors there).

A cold enough body, though, has a low enough vapor pressure that this isn't relevant even over cosmological timescales. That's why Europa can can have a stable icy surface. It's far enough from the Sun (and has a low enough albedo) that it's very very cold (about 100K), and at that temperature ice doesn't sublimate very much.

TLDR: a Ceres-sized ball of water could hold itself together, but only as long as it stayed water. But it wouldn't be able to. Either it'd be cold enough to freeze over at the surface, or hot enough to evaporate into vapor that would escape.

[1] https://en.wikipedia.org/wiki/Atmosphere#/media/File:Solar_s...

[littlestymaar]

Given that water gets lighter when cooling down right above its fusion temperature, and that ice is a pretty good insulator. You'd have liquid water below an ice crust for a lot of time. It would eventually freeze entirely and be slowly eaten by the Sun's radiations. But that would take a pretty long time (well on a human scale).

[rachofsunshine]

Yeah, that's why I specified freeze over and not freeze through, although without doing the math I'm pretty sure it'd still freeze through on solar system timescales without radioactive (as in Earth's own mantle's case) or tidal (Enceladus, Europa, possibly Triton and Ganymede) heating.

[littlestymaar]

Indeed, it will slowly freeze though and evaporate at the same time.