Who cooks vs what's cooked: the owner-ethnicity puzzle in LA's immigrant restaurants

Walk into Marouch in East Hollywood and you will eat some of the best Lebanese food in Los Angeles — fattoush, shawarma, makanek, the muhammara that regulars drive across town for. The family that runs it is Armenian. Walk into Nazelie’s in Glendale: a Lebanese menu, Beirut-rooted, Armenian-owned ^[1]. This is not a quirk. It is one of the most reliable patterns in immigrant Los Angeles — the cuisine on the plate and the ethnicity of the person who owns the restaurant routinely point in different directions — and any directory that tries to describe LA’s food honestly has to decide what to do about it.

The pattern, stated plainly

A restaurant is a format and a menu. An owner is a person with a community history. In LA those two things diverge constantly:

• Marouch, Nazelie’s, Sahag’s Basturma — Armenian-owned, Lebanese (or Syrian) menus. This is the Beirut-Armenian diaspora: Armenians who settled in Lebanon after 1915, cooked and ate Levantine food for generations, then emigrated again to Glendale and Hollywood. The Armenian part is who they are; the Lebanese part is what they cook ^[1].
• Komodo / Phorrito — a Vietnamese-Mexican menu (the “Phorrito,” pho in burrito form) run by an Indonesian-American chef. Neither half of the menu matches the owner’s heritage at all ^[2].
• LA’s independent donut shops — by most counts 80-90% are Cambodian-owned, and they sell American donuts: glazed, old-fashioneds, apple fritters, the pink box. There is nothing Cambodian on the menu ^[3].
• Korean-owned sushi bars — a large share of LA’s mid-tier Japanese sushi counters, especially outside Little Tokyo, are Korean-owned. The food is Japanese; the owners are Korean ^[4].
• Vietnamese-owned Chinese restaurants — the Hoa (ethnic-Chinese Vietnamese) diaspora runs many of LA’s Cantonese and seafood houses; here the owner’s heritage and the menu’s cuisine actually agree (the owner is ethnically Chinese), but the national label diverges — a “Vietnamese-owned Chinese restaurant” in casual speech is really a Chinese-from-Vietnam restaurant ^[4].

Why it happens

These divergences are not random; they come from how immigrants actually enter the American restaurant economy.

Refugee-economy entry points. When a refugee community arrives without capital or English, it does not invent a cuisine — it buys into whatever restaurant format has low barriers, cheap equipment, and a teachable routine. For Cambodians fleeing the Khmer Rouge, that format was the independent donut shop: Ted Ngoy trained at Winchell’s, bought Christy’s in La Habra in 1977, and built an empire by teaching the donut trade to relatives and other refugees ^[3]. Donuts were the door, not the destination.

The sponsorship-chain model. Once one family establishes a foothold, it sponsors more — co-signs leases, fronts equipment, trains the next arrivals, who in turn sponsor the next. Ngoy sponsored well over a hundred families; the pattern compounds until an entire trade is dominated by one community ^[3]. The same chain logic explains Korean ownership of Japanese sushi bars and Armenian ownership of Levantine restaurants: a community gets good at running a format, and newcomers from that community plug into it because the support network is already there.

Buying into established formats. It is often easier — and lower-risk — to take over an existing, profitable restaurant concept than to introduce an unfamiliar cuisine to American diners. An Armenian family in Glendale opening a Lebanese restaurant is opening a cuisine Angelenos already recognize and the family already cooks; an Armenian family opening an “Armenian restaurant” in 1985 faced a market that had never heard of zhingyalov hats. The path of least resistance is to cook what sells.

Layered diasporas. And sometimes the divergence is just history doing its work twice: the Beirut-Armenians, the Hoa Vietnamese, the Persian Jews — communities that were already a minority inside another country before they emigrated to LA. Their food is the food of the place they last lived; their identity is older than that.

How this directory handles it

Collapsing the two would break things in both directions. If we filed Marouch under “Armenian,” a diner searching for Lebanese food would never find one of the best Lebanese kitchens in the city — and we would misrepresent the menu. If we invented a “Cambodian donut shop” category, we would either erase the Cambodian story (by filing those shops as plain donut-shop) or absurdly relabel 85% of LA’s donuts as a foreign cuisine they aren’t.

So the model is two fields, doing two different jobs:

• cuisine = what is on the menu. Marouch is lebanese. Komodo is vietnamese (with a fusion flag). A Cambodian-owned corner shop is donut-shop. A Korean-owned sushi bar is japanese.
• heritage (separate field, plus generation where relevant) = the owner-community story. Marouch carries heritage: armenian and a note on the Beirut-Armenian diaspora. The Cambodian donut shops carry heritage: cambodian-american and generation: first|second, plus an editorial note: on the donut-shop slug itself describing the Ngoy lineage — rather than a cambodian-donut-shop cuisine slug, which would model ethnicity as if it were a service format ^[3].

The rule of thumb: taxonomy models format and menu; heritage is metadata about the people. When the two agree (a Oaxacan family running an Oaxacan restaurant), there is nothing to resolve. When they diverge, both facts get recorded, neither gets overwritten.

Why getting it right matters

Because both facts are true and both matter. The Lebanese food at Marouch is genuinely Lebanese — that is not a technicality, it is the point of going. And the reason it exists in East Hollywood, run by an Armenian family, is a story about genocide, displacement, Beirut, civil war, and a second migration to Glendale — and that story is worth telling too. The pink donut box is a real American object and a refugee-economy artifact of the Khmer Rouge survivors who own the shop. A directory that flattens these into one label has to pick which truth to keep, and it will pick wrong. Two fields is not bureaucratic hedging; it is the only way to be accurate about a city whose food is this layered.

Draft — Search session, 2026-05-11. Citation note (2026-05-12): this is a “big idea” cultural-note — the substance is Delicioso’s own taxonomy reasoning (the cuisine + heritage/generation two-field model and the LA examples that motivate it), which is appropriate for a cultural-note but is largely not externally sourced. Hardened what’s externally checkable: Marouch (1982, Serge & Sossi Brady — Resy) and the Ted Ngoy / Donut King lineage (Wikipedia, LAist). Founder review before publish: pin public coverage for Komodo / the “Phorrito”, the Korean-owned-sushi-bar pattern, and the Hoa-Vietnamese-Chinese-restaurant pattern, or keep them framed as observed patterns rather than sourced claims. owner_domains: [armenian, cambodian, korean] — confirm intended encoding for an editorial-frame piece.