Why do different mass definitions (Komar, ADM, Bondi) yield conflicting signs for the same solution?

This is a genuinely subtle question about how general relativity defines one of its most basic quantities. The short answer: the definitions don't actually conflict on *physical* solutions, but they measure fundamentally different things, so they can give opposite signs on *unphysical* or artificial spacetimes.

Why the definitions exist

In Newton's theory, mass is simple: it's a number you measure once. General relativity breaks this down because spacetime itself is dynamical. When you ask "what is the mass here?", you must specify *where* and *when* you're measuring it [AG-2025.03-718].

The three main definitions each pick a different vantage point:

• Komar mass: uses a symmetry (a Killing vector field—a direction in which the geometry stays the same) to define energy locally. Think of it as "the energy I can extract using this symmetry."
• ADM mass: measures total energy at spatial infinity on a spacelike slice, like stepping back far away and asking "how much does this whole thing weigh?" [AG-2025.09-142]
• Bondi mass: measures energy carried away by gravitational radiation escaping to *null* infinity (along light cones), so it decreases as waves escape [AG-2025.03-718]

When they disagree on sign

The practical answer is: they agree on real astrophysical solutions. The sign disagreements occur on pathological or deliberately constructed spacetimes that violate physical assumptions.

For instance, one can write down a metric that formally satisfies Einstein's equations but contains a "negative mass singularity"—a region with negative Komar mass [AG-2025.10-198]. However, such solutions are typically unphysical: they violate energy conditions (roughly, the requirement that matter doesn't have negative energy density in normal frames), or they're dynamically unstable (any small perturbation causes them to collapse or explode) [AG-2024.07-629]. General relativity itself "abhors negative masses" classically—it won't maintain them as stable solutions.

The deeper connection

Recent work shows that ADM mass and a properly generalized Komar energy are actually *equal* in realistic dynamical spacetimes, provided the geometry falls off fast enough at infinity [AG-2025.09-142]. The key insight is that you don't need a Killing vector to do this; you can use the lapse function (the coordinate time variable) instead, which exists everywhere. This unifies the definitions in a covariant way.

Similarly, ADM and BMS (Bondi-Metzner-Sachs) quantities are related through subtle connections in how you define conserved charges at different infinities [AG-2025.05-647].

The upshot

Different mass definitions don't conflict in nature—they're different questions asked of the same spacetime. Sign disagreements signal that you're working with an unphysical model (usually one missing real physical constraints like stability or reasonable matter content). When working with real gravitational systems, the definitions converge to the same positive value.