Let us assume a high-boiling non electrically conductive oil in a tungsten filament lamp, and compare this with a low pressure gas, which is the normal method (i.e. a vacuum but not a perfect one).
In the case of the oil, what would happen is like an element in an electric kettle. The element would get hot and heat the oil. That is to say, the molecules near the element would get hot by thermal conduction. This region would expand and move away from the element by a process of convection (as hot air rises) carrying heat energy around the liquid (as happens with a pot of liquid on a stove). Thus the oil will get hotter and keep the element cool.
However, a cool element does not emit light, you need to get it white hot.
There are few solids which do not melt at the temperature of the light bulb (which is why tungsten is used) and no liquids which do not boil or decompose well below that temperature.
It is, however, possible that the system would reach a temperature in thermal equilibrium where the glass of the bulb is losing heat to the surroundings at the same rate as the filament is supplying it. However it would not be giving out any light.
That is, you have constructed something equivalent to what is normally called an oil-filled electric radiator.
Would it last for ever, this light bulb consuming power and not giving any light?
The convection of the oil causes motion which will gradually wear away the filament as the Colorado River has worn away the Grand Canyon.
(There are other engineering issues such as the expansion of the oil which would burst the glass, but I can see ways in which this could be overcome by some complications to the design.)
In a normal light bulb there is only a tiny mass of material, so the effect of convection is likewise very tiny and the filament heats up rapidly to white hot. The main heat loss mechanism is radiation which is vastly less effective than liquid convection.
The filament eventually fails because at the high temperatures atoms of the metal actually vaporize, and can condense on the cooler surface of the glass, causing darkening. In a normal filament bulb a small pressure of gas inhibits this to some extent.
In the halogen bulb there is an element (e.g. iodine, a halogen) which reacts with the tungsten vapour at the lower temperature of the glass (quartz in this case) but decomposes in to its elements (tungsten and iodine) at the temperature of the filament, redepositing the metal atoms.
This answer was by Martin J Pitt, Chemical Engineer, Chemist, University Academic and Lecturer.PhD Chemical Engineering, Loughborough University Graduated 1985 Lives in The United Kingdom