Are you referring to a character's thoughts or his voice heard from his subjective POV? I assume the first.
Like @Ian suggested we often tend to go for the "voice over" approach. I guess because voice over sometimes also has such qualities. (In cases when one of the characters is also the narrator)
So if you don't have any other narration, this approach should work fairly well.
Though I would say that the style of the voice has to fit the context. Which means the story, it's character, what is said, other kinds of vocal expressions,…) So if there is also narration or the character is anything but soft and gentle, this won't work.
Basically you have a lot of influence in all three stages of the sound process: recording, editing and mixing.
While recording, as always, performance rules! Like @Utopia said if it is not there, you won't get it done by mixing. But for such cases as internal voices I would always try recording different microphones in multiple positions. Thus you have more choices and you can also work with combinations or try to fake some sort of stereo in order to widen the voice and thus clarify the difference.
While editing don't just clean it up. (You will though need to work a lotto remove all lip sounds and the likes) Try to be creative. E.g. have lines overlapping if the mood is more agitated or something like this.
While mixing think also in terms of dimension. So (as mentioned above) you can use different mics to create a stereo image and thus get the voice off the center only.
I would not go for too much reverb, if at all. I think this works for dream-like situations, but other than that, I'm not sure. But maybe that's taste.
This is just to give you some hints for directions to play with. In the end sound is relative and you need to find the right relation to all the other voices and sound elements as well as the story.
As it was mentioned by @Andy Lewis, on Enter the Void we had to find (at least) four different characteristics of dialogue. There's the normal on screen dialogue (which has it's own sound due to the director's wish). The subjective POV of the protagonist. His thoughts heard in between other lines. And some druggy inner voice. In the case of the thoughts our goal was to mix about three to four takes for one line. One is prominent and the other ones are a little less and edited in a special way that you's always hear at least two of the same lines at a time. In terms of mixing there is no reverb on it and they are mono on contrast to the wide POV voice. And they have less base and high end. Just to encourage some experiments…