Beirut is so well covered in the press that I can't really give much new information. Maybe I would just mention that our favourite meal was at Seza, in Mar Mikhael. Seza is an Armenian restaurant. While I know a few things from that cuisine, a lot was new to me. It was warm enough to sit on the porch, which is lovely.
They have a standard tasting menu but it contained too much we weren't into (=lots o' carbs), and missed a lot of things we wanted. In the end including way too much Arak even by North German standards we ended up spending 1 to 2 USD per person more than the menu for a YUGE amount of food.
a garlicky tomato salad, which added a nice brightness to the rest of the meal.
"itch", a bulgar salad with onions, tomato and various spices.
raw kibbeh with a pile of chopped herbs in the centre of the plate. Quite possibly the best version I've ever had, almost whipped consistency
chicken livers in pomegranate molasses, excellent version (I make this at home sometimes.)
hummus with awarma, which is basically lamb confit-ed in lamb tail fat. The lamb confit easily blew all other versions I've had out of the water. Consistency like well made carnitas.
A sausage dish. I thought it would be something else, but ended up being similar to sliced chorizo in a tomato sauce. OK.
fried fish -- cut into crosswise sections, super simple but perfectly executed. What the kids let us try was delicious. Served with brilliant garlicky fried potato cubes. (And I typically don't like much garlic..) They also made a portion of potatoes without the garlic for the kids, without us asking. Service!
Meatballs in a sour cherry sauce.
Manti, again a category killer. Best I've had.
Desserts were on the house. IIRC there was rice pudding, a baklava type thing and I forget what the third was. Needless to say we were pretty full that that point.
Verdict? All but maybe the sausage I would order again. Three of the dishes were the best versions I've had anywhere. How often does that happen?
Here's a blog with pics.
Summarizing, then, we assume that relational information is not subject to a corpus of utterance tokens upon which conformity has been defined by the paired utterance test.