Again I say, as someone who has lettered comics, the difference is purely cosmetic. Functionally speaking, thought bubbles are identical to first person narrative boxes.
Which is pretty much my point -- that it's just a matter of style. One style isn't superior to the other, just different. I don't object to first-person narrative boxes, but I don't see why their use in some comics should preclude the use of thought balloons in other comics -- any more than, say, the use of electronic music in some TV/movie scores should preclude the use of orchestral music in others, or the use of both in various combinations. I don't see any reason why both stylistic options shouldn't remain on the table. Having more options to choose from is good.
From a practical standpoint, I think boxes work better because they allow a bit more freedom when it comes to layout. It's easier to shove a box up in the corner or along the bottom of a panel without obstructing the art, while thought bubbles can very quickly start to clutter things up, especially if there's a combination of spoken dialogue and inner monologue. As a letterer, clutter is your mortal enemy.
Isn't a lot of comics art designed with balloon placement in mind, though? I've seen lots of page thumbnails with the balloon placement roughed out from the start. And speech or thought balloons can be right on the top of a panel with a long "tail."
Plus, there can be occasions when you don't want a narrative box to be shoved off to the side. Sometimes their placement within the frame is important too -- for instance, the second illustration in
this column, where one character's "internal monologue" text boxes are directly superimposed over another character's speech balloons, obscuring the words, in order to show that he's too absorbed with his own thoughts to hear what the other person is saying. Or in scenes where Deadpool is having a spoken conversation with one of his inner voices or with the narrator, you need to put the thought/narration boxes in between his speech balloons to show the sequence of the conversation.
The rules about scene changes to switch POV also generally apply here the same way they do in prose, though if you have to do it, it's a lot easier to pull off again, thanks to the use of distinctive colours & fonts. Still, it's fairly rare, even in comics to cross POV like that and of the times I've seen it in use, it's usually done as a gimmick. An example that leaps to mind is the 'Batman/Superman' book from about 10 years back. It would have all of Bruce's inner thoughts down one side of the page and Clark's down the other, intentionally juxtaposing each other's views and mindset.
Yeah, I've read some of those. It got annoying fairly quickly.
My theory: The issue is that they're not diegetic to the same extent word balloons (or telepathy balloons) are. With a word balloon, that's actually exactly what a character in-story is saying. But people don't actually articulate their thinking in neat, organized sentences, not usually -- we understand what's in a thought balloon is an interpretation, from thought to language, for the audience's sake. It's not what's *really* happening, in-story.
The same is true of italicized internal monologue in prose, though. I'm okay with the convention for that reason. Also, as I said earlier, sometimes I've wished that certain stilted dialogue passages in older comics had been put in thought balloons instead of speech balloons, because nobody would ever actually say such things aloud, and casting it as a "translation" of inner thoughts into words for the reader's benefit is more plausible.
There's a similar difference on a visual level. We don't think about this much, but a word balloon isn't just a representation of what a character is saying. It represents the actual *physical sound* of those words. That's why we intuitively understand what it means when a word balloon is giant-sized, or when one's shape is rugged. Why we'd think it's weird if the tail of the balloon aimed not at a character's head (where the mouth is) but their foot or hand. Thought balloons, on the other hand... well, nobody's thoughts actually float or emanate out of their heads, the way a thought balloon does. The graphic depiction doesn't correspond to anything in-story.
So you have word balloons and thought balloons, these two storytelling tools that look so similar and often exist side by side in the same panel. However, in not one but two ways they exist on different diegetic "levels." My theory is that this contradiction creates a mild cognitive dissonance, and that this dissonance is why some people perceive thought balloons as awkward.
Interesting thought... but couldn't you say that thought boxes and omniscient-narration boxes exist on different diegetic levels too, since only one is actually coming from the characters in the frame? Not a problem if your story is first-person narrated, but not all stories are.
