|Roast barley, earlier.|
Of course, "historically" there's a great deal of truth in that. Naturally, it depends what you mean by "historically", but we'll let that stand. But what about non-historically? Or, as we might say, "now".
How many beers are there that we can be sure the brewers are calling stouts and list their ingredients?
Let's have a quick look...
29 stouts, 21 with roast barley.
15 imperial stouts, 13 with roast barley.
And the "porters"?
23, 5 with roast barley.
There's a handful where it's hard to tell if they're being thought of as a stout or a porter. And a couple explicitly referred to as sort of hybrids (both contain roast barley, FWIW).
So let's consider the incidence of roast barley in stouts v. porters.
Imperial Stouts: 87%
Other Stouts: 72%
All stouts: 77%
So, you pick up a bottle conditioned stout, it's more than three times as likely to have been made with roast barley than a random porter. That's to say, the majority of stouts are made with it (based on this selection), while the majority of porters exclude roast barley.
Now you might argue, as does Mr Cornell, that this distinction has no "historic validity", and you'd have a good point. You might even choose to stress that here we have "beers being called porter" rather than the true descendants of historical porters. But you know, history tells us mostly about change. And clearly, now isn't entirely like then was.
A diachronic approach to the "porter" / roast barley question? Or a synchronic one? As someone who brews beer nowadays, it's quite clear:
What's the difference between most modern, British stouts and porters? Well for one thing, and much more often than not, the use of roast barley.