You seem to believe that if a nation doesn't have a rigorously defined ontology, that makes it completely arbitrary. Which is a very strange position to take. Political parties, social classes, economies, and even
consciousness run into the same problems. But all of those phenomena arise naturally and we assume they serve some kind of purpose. The fact that nationalism is expressed through the state only means that the state is useful for expressing it, not that it is
required to (for instance, people of different communities can use 'Armenia' or 'Turkey' as a quick reference, instead of going through a list of all their cultural idiosyncrasies to compare themselves).
There's a difference between "completely arbitrary" and merely contingent. I don't believe that German identity is arbitrary; there's a reason why, for example, it was successful in Saxony but not in Bohemia. But I don't believe that German identity, as we understand is historically necessary; the historical forces that placed Austria outside of "Germany" but Bavaria within it are bound up in too many variables to see it as an inevitably or even particularly likely outcome. It's simply the outcome we got- or, even more strictly, that we have for now.
The state, on the other hand,
does require the nation to be something definite and independently-existing, at least so far as those claims to nationhood are based in an ethnic or cultural identity, and not in the constitutional continuity of the state itself. For the state to represent and embody the nation, the nation must be something capable of being represented and embodied. It's not sufficient for it to be just a cluster of reference points. Moreover, to maintain an authentically national character, the state requires that the nation be embodied exclusively by that state: two Frances or two Germanies are not sustainable, not without remaking one or both of the national fragments. Nobody really thinks of Taiwan as "China" any more, whatever the constitutional and diplomatic formalities. Nationhood is not sustainable so far as it is just
one of the characteristics of a national state, it is be the defining and fundamental characteristics of that state, or the state is revealed as an institution that exists for its own sake. It's not enough for nationality to be something
about the state, it has to be what the state
is. Look at the recent legislation in Israel: what was that, if not an attempt to more rigorously define the character of the Israeli nation, and in doing so to affirm and clarify the "national" character of the Israeli state?
To the extent that language and culture become merely customs, merely the norms of public life, the foundation of the state is the state itself, is the state's own constitutional continuity. This probably pays some part in the lack of nationalist zeal in contemporary Germans, who try their very darndest to understood "Germany" in terms of the continuity of the Bundsrepublik, and not as tracing back to the misty Teutonic forests. It may be why modern nationalism tends to take an authoritarian slant in democracies, because it is necessary to ground the state in "the nation" only to the extent that the populace struggle to identify themselves with the state itself. The more open and democratic a state's institutions, the less necessary it is to imagine "the nation" in anything but civic terms. It stops being "the nation" so much as it becomes "the republic". Which, I've suggested previously, would not be a bad thing.
You also suggest the existence of nationless states, or nations that exist outside of states, but what is the process by which tribes or confederations or ethnic groups are made into "nations", if not through the state? To talk of Armenians or Kurds as a "nation" is to express an aspiration to statehood- or at least, to some form of self-government, to "national" institutions within the state, if not independent of all other states. There are more Gujurati than there are Poles, but we don't talk about a "Gujurati nation", because there is no clearly or commonly expressed aspiration among Gujurati for independent statehood. We might talk of a "Sikh nation", but doing so takes on a distinctly political cast, carries immediate connotations of Sikh separatism and the aspiration for an independent statehood. While nationality may not always be defined in explicit reference to the state, it is very generally reserved for peoples who possess, or might plausibly acquire, independent states, or significant degrees of self-government within existing states.
And the state isn't always involved - national identity can last for decades in expatriate communities, which I presume you would call delusional?
However, diaspora identities don't always have a lot to do with "homeland" identities. Irish-Americans and Irish-Irish don't have as many common reference points as the former might like to imagine. Experiences diverge, and what you're left with is shared history, not a shared present. How many Irish-American families take a clear position on the Fianna Fail/Fine Gael split, which was one of the primary reference points in Irish public life through the twentieth century? The nation as an eternal now falls very swiftly apart when it strays from the homeland. The persistence of diaspora identities and their construction as "national" identities is interesting, but too complex and often too specific to prop up the claims of homeland-nationalists.
