Recessed Lighting - Thermal Cut-Off or Electrical Short

I've got a fairly standard covered, recessed lighting fixture in an upstairs bathroom, where there are in fact, two of these fixtures on a single circuit.



Although both fixtures are equipped with the correct watt light bulb (Max: 75W), periodically I have noted that the bulb in one fixture or the other will seemingly fail, but then after a period of time, will illuminate again.

This sort of not-easily-explained electrical "behavior" always seems to freak my wife out, since she seemingly never met a risk (real or imagined) that she didn't want to obsess about (evidenced by a prior posting on Outdoor Lighting, to ward off imaginary burglars she felt were lurking in the backyard . . .). And for the record I try to accommodate some of that paranoia, up to a point (. . . I only acknowledge that publicly, since she won't ever see this due to concerns about the NSA spying on her on-line . . .).

Consequently, my wife is convinced that one or both light fixtures have a dangerous "shortage" in them (I don't correct that use of terminology, only because it's always entertaining to hear her say it). My belief is that both fixtures must have some sort of thermal sensor built into them, which shuts off the flow of current to the bulb, when the temperature within the fixtures reaches a certain level. I'm happy if that's the case, because the predominant mass of the fixture does extend up into the attic, presumably anchored to one of the ceiling joists.

I've noted that the fixtures don't shut off that frequently, and typically only after they've been left on for an extended period of time (2-3 hours). Since we've correctly concluded that there's nothing wrong with the bulbs, or the switch, and eliminated the likelihood of gremlins, elves, and even prank-playing burglars (thanks to the outdoor lighting now . . .), does this theory about the integrated thermal limiter sensor in the fixture sound right?