Let us consider light. In this (the OP's) universe, let light be a massless particle with energy but no charge... much like in our own universe.
However, let us also consider anti-light. In this universe, anti-light is also a massless particle with energy but no charge. However, it is also the antiparticle of light.
When a particle of light and a particle of anti-light meet, they cancel one-another: When a particle of light meets a particle of anti-light, their energies are summed. If they had equal energy, then there would be zero result. If one particle had more energy, then the energy of the other would reduce the energy of the greater to the sum of the two.
As an example, say that we have two particles, a light particle with +5 energy and an anti-light particle with -6 energy. When they collide, an anti-light particle with -1 energy will continue on in the same direction as the original -6 anti-light particle.
Let us also consider optical temperature. In this universe, let optical temperature be a measurable property which is expressed as a signed real number, ranging from some negative value, through zero to some positive value. As optical temperature decreases, the particles vibrate less in one dimension and below zero, begin to vibrate more in a dimension at right-angles to the first. However, as objects lose thermal energy and physical temperature, their optical temperature tends toward the positive, not toward zero.
Now, let us consider the optical temperature of the universe and the objects within it. Let the universe's background optical temperature approach some value +n. In places which are warmer, the optical temperature will be lower, not higher.
If we carry this trend onwards, at some point, hotter objects will attain zero, then negative optical temperature, until they reach the value -n.
Now, let us consider the phenomenon of black-body radiation. Objects with a non-zero optical temperature emit radiation. Positive optical temperature objects emit positive light. Negative optical temperature objects emit negative anti-light. The further from zero the value, the more energetic the emissions, positive or negative.
Now we need to consider the structure of the universe. The universe is mostly empty space. Cold inert matter which is far from stars will have an optical temperature which approaches +n. Such matter will emit light. Planets which are closer to stars will be warmer, having a positive optical temperature considerably less than +n. They will not emit as much light, and like black-body radiation in our own universe will appear to be a different colour.
Stars, on the other hand, are much more energetic, and have negative optical temperatures approaching -n. They will emit anti-light, which will cancel out the positive light emitted from the less energetic parts of the universe with positive optical temperature.
So, why have I said that cold and light are positive, while hot and anti-light are negative? Because of what sentient creatures see and describe. The sentient creatures on a particular world in this universe can only see light, not anti-light.
That means that they will see the light from all the cold things reflecting off the warmer things, but they won't see the anti-light from the hot things at all. In fact, being close to something hot enough will mean that the anti-light that the hot thing emits will cancel out the light from all the cold things nearby... and will therefore cause darkness.
An interesting side-effect of this light/anti-light duality is that there may exist species which can only see anti-light, and they would consider stars to be bright, and the rest of the universe to be dark.
Additionally, there may exist species which can see and distinguish between both light and anti-light. Everything would likely be visible to such a species, but light and anti-light would be different colours in a way that those that can see only light or only anti-light wouldn't recognise.