There are several ways to decide this question, all fairly easy...
To reflect a star's light in such a manner as to emulate the entire output of the star itself, think of how colossal the artificial reflecting surface would have to be: it would have to be star-sized and in very close proximity to the emitter. What's the point of constructing such a thing?! Occam's Razor says if it looks like a natural star, it is one.
Any reflector which only diverts a portion of a star's light towards Earth would be detectable as not a real star, because it dims the star out of proportion with its expected output based on its spectral type and distance estimators. Thick interstellar dust along the line of sight could not be blamed for this dimming, because dust also alters the star's spectrum through reddening.
We do see reflected starlight in some circumstances - reflection nebulae and light echoes. However these are not point sources - their extended nature is obvious - so they do not mimic a star and could never be confused for one. They also fall under the same detectability as case 2): only a portion of the light is reflected.
We can't tell if a light signal has been reflected or if it is still in its original direction. The same is true about refraction. If we can see the same thing twice then that tells us that one or both of them is reflected or refracted. This is why people think gravitational lensing takes place, because the same object (by accurate measurement of its colour spectrum and other properties) is seen in two places, either side of the direction where the centre of the gravitational lens is thought to be.
There are several ways to decide this question, all fairly easy...
To reflect a star's light in such a manner as to emulate the entire output of the star itself, think of how colossal the artificial reflecting surface would have to be: it would have to be star-sized and in very close proximity to the emitter. What's the point of constructing such a thing?! Occam's Razor says if it looks like a natural star, it is one.
Any reflector which only diverts a portion of a star's light towards Earth would be detectable as not a real star, because it dims the star out of proportion with its expected output based on its spectral type and distance estimators. Thick interstellar dust along the line of sight could not be blamed for this dimming, because dust also alters the star's spectrum through reddening.
We do see reflected starlight in some circumstances - reflection nebulae and light echoes. However these are not point sources - their extended nature is obvious - so they do not mimic a star and could never be confused for one. They also fall under the same detectability as case 2): only a portion of the light is reflected.
To excellent Ray's answer I would add that if the reflected stellar light passes through an atmosphere, one can detect absorption lines in the spectrum that aren't there in the star spectrum. These lines show for example that the atmosphere temperature is very different from the star surface temperature.
If the iight is reflected from a hard surface, like for the Moon, provided one has the original star spectrum, one can detect a difference of the spectrum overall shape wrt to the star one, because different colours are differently absorbed or diffused by a hard surface.
What you want to look for is some change to the incoming radiation that is effected by the reflector itself. This often means that you must recognize what must be the originating source of the radiation and find that the radiation you see has been altered by the reflector. Unfortunately this is very difficult to determine unequivolcally. You may have to know something about the reflector in advance. For example, we infer that the planets shine by reflected sunlight because they are too small to generate the light and solar spectrum we observe on their own. Thus, reasoning must be employed to infer whether or not a particular source of light originates within the object or is reflected by the object.
With regard to Ray's first point, astronomical reflecting telescopes do a very good job of reflecting starlight including the star's spectrum so that I doubt that one could discern whether one was looking directly at the star through a refracting telescope or viewing reflected starlight from the mirror of a reflecting telescope.
And the image in a reflecting telescope will be inverted! With regard to polarized light you don't know a priori the cause of the polarization but you are right it's at least an indicator of reflected light. Thanks for pointing it out!
An interesting question. Analysis of light from distant stars is a key knowledge of other worlds, other planetary systems, exoplanets. Considering current research problems important for science, I think that the question concerns particularly important issues. I believe that research on this topic should be continued. There was an interesting discussion. Greetings