> it should be encouraged at every step, though never forced
I think you'll find that most of your "wacky libertarian" buddies don't see the distinction between these two methods.
> Probably, the answer is somewhere in the middle, and I don't claim to know where it is.
I'd recommend reading this book: http://www.amazon.com/Talking-Strangers-Anxieties-Citizenshi... One of the things it helps do is explain how and why we should negotiate the middle.
Democracy isn't about accountability or critical publics, though I'll concede the third point about tools for discussion, which we don't have enough of, and which are of fairly pitiful quality in this day and age. (That's going to be a contentious claim, isn't it? Damn. I'll footnote my complaint as [1] then.) Accountability and critical publics are, I agree, necessary but they're nowhere near sufficient.
I do not think that it is working particularly well in any country. No country I know of has a population dedicated towards an increased participation in their own governance. At best, they ask a class of well-to-do persons to handle their governance on their behalf.
But I also think a well-done democracy is actually an ideal worth pursuing. It is not enough to imagine democracy as the least tyrannical, most tolerant state of government: it must be something positively defined to be recognized as an ambition. For that, I turn to a book published in 1916, John Dewey's Democracy and Education.
"A democracy is more than a form of government; it is primarily a mode of associated living, of conjoint communicated experience. The extension in space of the number of individuals who participate in an interest so that each has to refer his own action to that of others, and to consider the action of others to give point and direction to his own, is equivalent to the breaking down of those barriers of class, race, and national territory which kept men from perceiving the full import of their activity." [2]
To me, a well-done democracy is one in which the overwhelming majority of the population (for those who want a number, let's call it 70 to 80 percent) are actively participant in their governance. One possible way to measure active participation is the drafting of legislation. (Voting is passive participation; a head count, to paraphrase Einstein [3], requires only a spine. It does not require a large brain.) Sure, this requires free flow of information, critical thinking, and tools for discussion, but you begin to recognize that having gcc, gdb, and a decent CS education is not sufficient to write an operating system from scratch: you have to actually put code to computer. For me, such a goal necessitates a reconstructed educational system. Perhaps I am wrong about that, but if I thought that, I wouldn't be trying to come up with one. :)
But wait, there's more. It isn't sufficient to have heavy participation. It has to be maintained. Such sustained participation isn't possible if some particular subset or coalition gains an upper hand: and yet that is exactly how any kind of legislation can ever be passed. What's needed here is ways to guarantee that underdog groups always have their voices acknowledged, heard, engaged with, and accounted for. An effort should be made to keep them from losing every vote; such an effort, I expect, will inevitably lead to a fracturing of any underdog group whose interests actually non-beneficial. That's an unproven guess, but it is my expectation. You can read more about this particular issue in the place my thinking on it was inspired from, Danielle Allen's Talking to Strangers [4].
And lastly, there needs to be a kind of standard, a set of shared values that can undergird the very notion of a democracy. Why is it ideal that everyone has a voice? The core behind it is an egalitarianism: the bedrock belief that no one is intrinsically worth more than another. This egalitarianism has often been expressed as an equality of outcome (proportionate representation), or an equality of opportunity (non-discrimination), or an equality of process (everyone jumps through the same hoops). I favor Amartya Sen's equality of autonomy [5] which leads into the capabilities approach that he developed alongside Martha Nussbaum [6]. I like the capabilities approach because it avoids many of the pitfalls that previous arguments have run into: you can't stymie its analysis by moving macroscopically to statistical claims or microscopically to anecdote: its equally applicable to both cases. It's on bases like this that rule of law can be built.
It's important, I feel, to aspire towards democracy rather than resign ourselves to it. Such resignation makes room for a willingness to trade in democratic powers for short-term gratuity, such as the ever popular "liberty for security" but also things like "education for solvency" or "inclusiveness for comfort". Even if we misstep or are embattled, we do need to believe that there's a worthwhile goal to shoot for. Otherwise it's a constant battle of attrition for rights until there are none left, because we can't positively assert them on their own merits.
[1] Our communication tools have not been developed with community or collaboration in mind. This is particularly notable with Engelbart's passing, but it's always worth noting how very exceptional it is every time a decent "social" space like Github or Discourse comes out. Our tooling has focused almost exclusively on one-to-one communication, rarely even bothering to consider how three or more people could interact. This might be simple laziness, but what it really says to me is how inconsequential the notion of existing in a community is for us as a society: there's no itch to be scratched by good tools for discussion. The itch isn't there. And for those few who have it and scratch it, there appears to be no market.
[2] http://etext.lib.virginia.edu/etcbin/toccer-new2?id=DewDemo....
[3] The original quote is, "He who joyfully marches to music in rank and file has already earned my contempt. He has been given a large brain by mistake, since for him a spinal cord would suffice."
[4a] http://press.uchicago.edu/Misc/Chicago/014665in.html
[4b] http://www.amazon.com/Talking-Strangers-Anxieties-Citizenshi...
[5] http://en.wikipedia.org/wiki/Equality_of_autonomy
[6a] http://en.wikipedia.org/wiki/Capability_approach
[6b] http://people.wku.edu/jan.garrett/ethics/nussbaum.htm
[6c] http://crookedtimber.org/2011/08/29/martha-nussbaums-creatin...
[6d] http://www.amazon.com/Creating-Capabilities-Human-Developmen...
Nothing about the phrase actually requires the tyranny be remotely permanent, and nothing about the concept of tyranny requires it either. Indeed, it's actually more poignant that the tyranny is both ephemeral and, more importantly, inevitable. This is the real issue. It's easy to comprehend the issue by looking at semi-permanent states [1], like Marxist class divisions or political polarization, but those are consequences.
When you place something to a vote and fail to achieve consensus, you have a majority whose will is being enacted upon a minority. This is commonly celebrated by moronic catchphrases such as, "the right side of history," but you still have a winner and a loser. And when the ephemeral majority seems to crystallize from the aether into an unmistakable demographic of some kind, you really have to wonder about this "emergent majority" concept.
The fundamental American problem is that it is incapable of trusting its government. (I can't speak to other governments; I don't live under those.) This lack of trust is well-founded, but that doesn't make it any less crippling. I said below that the appropriate counter to the tyrannical majority is a strong executive. The executive is capable of saying, "No," to a popular opinion in a way that the legislature cannot. (Indeed, this is a role Americans have shoved onto the judiciary, which means we can't actually say no until someone files a lawsuit.) The interplay between the legislature and the executive (the vote, the veto, the super-majority overrule, the refusal to enforce, the impeachment) is exactly the solution being proposed.
None of it works without an executive willing to exercise its options. See Obama's refusal to enforce immigration laws, for instance. (For the sake of this "academic debate", let's presume that he could enforce them to the letter if he wanted to.) That is a check against tyranny of the majority. So naturally, well... see for yourself [2]. This isn't meant as an anti-Republican statement; it's just the most recent example I could bring to mind off the top of my head.
The trust part comes back when you ask, "Why doesn't a minority group tyrannized by a majority group just leave?" You know, secede. This comes back to the emergent majority you spoke of: if you trust the majority to, you know, not be dicks about it [3]. To not always be the majority, then secession isn't necessary. You can expect that the issues will change and you'll be in the majority. Of course, what's happened with the polarized political class in America is that you stick around because you figure you can win next time.
This is a problem. It is the exact opposite of what you need to make sure that a majority is an emergent majority. Instead of melting back into an indeterminate goo and forming new majority/minority groups in each vote, you have battle lines being drawn for a budding civil war. And that's as far as I've figured out.
[1] "State" as in "state of being", not as in "nation-state". My thesaurus powers have failed me.
[2] http://judiciary.house.gov/index.cfm/2015/3/at-the-flick-of-...
[3] Book recommendation: http://www.amazon.com/Talking-Strangers-Anxieties-Citizenshi...