The bar in the Hotel Meurice was full of weighty African diplomats sinking whisky and muttering into mobile phones. In the lobby, a blonde model stalked about like a praying mantis, preceded by photographers dragging each other backwards in her path. And any minute now, I thought, as we sat in front of gilded mirrors which multiplied the fortunate of Paris like a godmother's wand, Rémi Krug is finally going to lose his temper with me and chuck me out of here.

It's strange how much passion a few simple questions can evoke. But before I tell you which ones, a little background. Champagne (Château Thierry to Reims to well beyond Troyes, via a thousand scattered vineyard fragments) is a much larger area than most people realise. It lies in the wine world's far north, where a few twists of the geological screwdriver, the topographical spanner or the microclimatic allen key can change a no-hope vineyard into a great one. It has 15,000 growers, all of whom know their own vineyards as well as they know their own feet. And it has just one AOC: Champagne.

My questions to Rémi Krug concerned the fundamental … well, what do we call it? Dogma? Article of faith? Fact? Myth? Anyway: the assertion (to be neutral) that the greatest Champagne is created by blending fruit drawn from the entire region, hence the one-AOC-fits-all simplicity. I put it to Rémi that this assertion suited large companies but was uncomfortable for growers, since it was tantamount to saying that they could never (stuck with their few hectares of vines) make great Champagne. I innocently wondered whether it wasn't counter-intuitive, especially given the proximity of Burgundy to Champagne where the same grape varieties, bar Meunier, reflect vineyard differences with what Burgundophiles consider rapturous fidelity. Why should Pinot Noir and Chardonnay do the terroir tango near Dijon, but fall over their feet a little further up the road, near Troyes or Reims? Wasn't the truth that Champagne has evolved to help the large companies create and market their blends most profitably and successfully, with the famous assertion a post hoc ergo propter hoc justification to ensure nothing ever changed? And hadn't the nuances of Champagne's vineyards been obliterated in the process?

No, no, no! cried Rémi. All nonsense. Far from being some kind of a large-company conspiracy, the dominance of blends in Champagne was an evolutionary response to the inadequacies of the single-vineyard wines which had preceded them. And the fact that Champagne was a sparkling wine - in other words one whose aromatic and flavour profile is in part built by a series of winery interventions - made the analogy with Burgundy inadmissible. The aesthetic desiderata were different. Case dismissed.

But what about Clos du Mesnil? It was, insisted Rémi, the very special exception which proved the Grande Cuvée rule. The sad fact was that most vineyards in Champagne could not produce 'complete' Champagne; Clos du Mesnil was one of the few rare spots which could. I see. But didn't blending produce large quantities of consistent but bland Champagne? Wouldn't more emphasis on vineyard origin give us more diverse and more intriguing Champagne? On the contrary, Rémi shot back, it was only thanks to the skill of the blenders that so much Champagne was as good as it was.

And if I was insinuating that Grande Cuvée was bland … Not at all, cher Rémi, quite the contrary; my heart leaps whenever anyone offers me a glass. My point, though, is that great blends occupy a certain zone within Champagne's overall flavour spectrum, and great single-vineyard Champagnes occupy another. That other can surprise, disarm, disconcert. We're not used to it; we're used to all that mellow, creamy, biscuity blended stuff. Single-vineyard Champagnes may well seem a bit spiky, singular, challenging and over-characterful. Just like real Burgundy did after the years when most of it contained dollops of Tlemcen.

Actually I didn't say the last bit, since I would certainly have found myself on my bottom on the trottoir of the Rue de Rivoli. We agreed to differ, shared a convivial bottle of Grande Cuvée, and parted amicably, though I suspect I'm not on Rémi's Christmas card list any more.

I'm not, let me stress, anti-blend; the Roederer 1997 I drank last week was all that anyone could want from a glass of Champagne. My point, though, is that alternatives to blended Champagne barely exist as yet; they are an unexplored continent. Until more growers, over several generations, get to use their best fruit in Champagnes which reflect the nuances of terroir truthfully, and until our expectations of Champagne broaden to include more character than is deemed acceptable by most blenders, then we will never know the true potential of Champagne. Which would be a shame.