In BrE the two words are distinguished. Capitol doesn’t come up much (and when it does it most often refers to the one in Rome) but the final vowel has the same value as in pot or option. It may be that BrE needs a pronunciation distinction because the word is comparatively rare, and if there was no distinction most listeners would understand capital, which could make for confusion.
I grew up in the northeastern US and since grammar school we were taught that capital and capitol were homonyms, same as principle and principal (the latter being the head of the school and therefore, spelling-wise, your ‘pal’).
Really? It doesn’t matter if it changes the meaning? I find that very hard to believe.
Whereas I have no trouble believing someone would lose for giving what is objectively a wrong answer.
I have a feeling this will not be entirely well received, but I will say it anyway. 
I am sure that there are some people in the United States who pronounce “capital” differently than “capitol” due to regional variations or personal speech characteristics; however, the vast majority of folks that think they do, actually don’t in their normal speech. They will say them differently when specifically queried about it, but if you were able to record them for a year and find all utterances of these words, they would be homophones.
After being presented with this recorded history, they still would not believe that they pronounce them the same.
I read a study/article about this phenomenon a while back. If I remember correctly, it was basically a function of how written vs. spoken language works. In your brain, you perceive the difference of these words, but our voices are much more lazy than our brains, so the difference that we believe exists does not actually present itself in speech. Unfortunately, I was unable to find the article.
I think that’ll be well received. It makes complete sense. Most people are appalled to hear a recording of themselves. So many false starts, half-sentences, umms, ers and other verbal tics. All of which are absent in their internal “transcript” of what they think they just said.
If tongues weren’t lazy we wouldn’t have things like the city “Louisville” pronounced by the locals as, roughly, “uh uh” with the second “uh” even more schwa-like than the first.
Not at all. I’ve mentioned the same thing before. Though it’s not just laziness. It’s also just paying attention vs. not paying attention. And context: how you say a word by itself is often different from how you say it in a sentence.
Linguists try to get past this by not actually telling you what words they are checking you for, and having you repeat sentences back fast enough that you don’t have time to think about what the words are. Only after that do they specifically ask you about the words, so they can see what you think you do.
Yes. Thank you.
This happens all the time on this board, and I usually want to respond as you have, but I usually don’t have the time.
I don’t know if it applies to words, but at least for sounds, it goes the other way too: someone who thinks the sounds in two words are identical may pronounce them differently. I can’t tell the difference between “dark” and “light” l or between aspirated and unaspirated t, but I’ve been told that people with my dialect reliably pronounce one or the other depending on the word.
I would find it entirely plausible that I might even distinguish between those pairs in what would otherwise be homophobes without being able to tell the difference myself.
I pronounce both as ka-piddle, with an emphasis on the “ka”, and “ka” pronounced like “cat” without the “t”. Raised in western Washington State.
Me too. Though capitol is not used to refer to a legislature (at least in the US), but the building that houses it. It’s akin to “house,” or “chamber.”