A few months ago, I got into the habit of asking people what they thought New York’s defining sandwich was. Most of them in one particular group answered “bacon, egg, and cheese.” A respectable answer. A delicious answer. And then I’d ask, with a kind of bait-and-switch energy I’m not proud of, “What about a bagel with lox and a schmear?” Some people lit up. Others looked at me like I’d just made up a word. What is that? one person actually asked, in earnest.

That’s a real divide, and it surprised me more than it probably should have. I’ve spent years moving between these worlds — the bagel-with-lox-and-a-schmear sandwich isn’t exactly obscure; it’s one of the most iconic New York foods in the canonical outside-looking-in image of the city. Woody Allen movies, Nora Ephron essays, half the postcards in any tourist shop. So, encountering people who genuinely had no idea what I was talking about wasn’t trivial; it was the discovery that two New Yorks I’d assumed were at least vaguely aware of each other were, in fact, much more separated than I’d realized. And the divide isn’t about ethnicity exactly, and it’s not about wealth, either — both groups had people in every income bracket I asked. It’s not even really about borough. The thing it’s about is harder to name, but I think it’s actually about speech communities, in the precise sense that Dell Hymes meant the term: groups defined less by who’s in them than by their shared norms for interpreting how people talk. The sandwiches are a tell. The deeper diagnostic is what you say when you order one.

(I should disclose, before going further, that I’m a vegetarian and therefore eat approximately none of the sandwiches under discussion. Make of that what you will. My best guess is that it puts me in a fourth speech community I’m not remotely qualified to describe.)

What I want to argue, gently and inconclusively, is that New York is best understood as several speech communities sitting on top of each other, sharing subway lines and barely overlapping in the way they talk. And the European theory we have for explaining cities like ours — the “multiethnolect” frame — was built on cities where one consolidated youth variety emerged out of immigration. New York didn’t do that. New York produced parallel varieties. The reason has to do with geography, immigration history, and — one can at least consider — sandwiches.

The Lox-and-Schmear New York

Start with the older, recessive variety. This is the New York of Russ & Daughters, of mid-century Brooklyn, of fuhgeddaboudit — the variety academic linguists usually call “White ethnic NYCE” and that everyone else calls Brooklynese, even when it’s from the Bronx. Its features are some of the most famous in American English: non-rhoticity, where fourth floor is realized as $\text{[fɔəθ flɔə]}$; BOUGHT-raising, with coffee and talk taking $\text{[ʊə]}$ rather than $\text{[ɔ]}$; the genuinely complicated short-a split, where bad $\text{[bɛəd]}$ and can (the noun) $\text{[kɛən]}$ are tense but bat $\text{[bæt]}$ and can (the verb) $\text{[kæn]}$ are lax — one of the most intricate vowel systems in any English variety; dental stops for $\text{/θ/}$ and $\text{/ð/}$ turning thirty-third into $\text{[tɔɪɾi tɔɪd]}$; a lexicon shot through with Yiddish (schlep, schmear, kvetch) and Italian (agita, gabagool). William Labov essentially invented modern sociolinguistics by walking into Saks, Macy’s, and S. Klein’s and asking clerks where the fourth floor was, just to hear how they said the r. (They mostly didn’t.)

This is the variety that several recent studies — Kara Becker’s work on the Lower East Side, Michael Newman’s New York City English — have found to be in retreat. Young White New Yorkers are pronouncing their r’s. They’ve simplified the short-a split into a more general before-nasals pattern. The raised vowel in bought is reversing in apparent time. The dialect that defined the city in mid-century film and television is now mostly performed, not produced — closer to a costume than a coat.

The features aren’t dying. They’re being redistributed.

The journalism on this has tended toward obituary mode: the New York accent is dying, fuhgeddaboudit, etc. The obituary writers are largely jousting at windmills. The features aren’t dying. They’re being redistributed.

The Chopped-Cheese New York

Here is the variety that the obituary-mode journalism keeps missing: While young White New Yorkers were quietly dropping their non-rhoticity, Patrick-André Mather’s 2012 replication of the Labov department-store study found that non-rhoticity was not declining among Black New Yorkers — it was being recoded as a Black NYC feature. Meanwhile, Renée Blake and Cara Shousterman were documenting how second-generation West Indian New Yorkers had built a distinctive variety of English, combining AAVE features (copula deletion, habitual be, t/d-deletion) with Caribbean prosodic and lexical retentions. Michael Newman and Peter Slomanson were showing that what they called “New York Latino English” — apical $\text{/l/}$ from Spanish substrate (clear $\text{[l]}$ even in coda position, where Anglo English has dark $\text{[ɫ]}$), syllable-timed rhythm, post-vocalic consonant lenition — was a focused variety in its own right, with high school peer groups (hip-hoppers, skaters, geeks) determining how much of it any given speaker used. And in the most striking recent finding, Newman, Bill Haddican, and their collaborators have shown that young Latino and Black New Yorkers are now leading several ongoing changes in the New York vowel system. The vernacular leadership has transferred.

This is the New York of the chopped cheese — a sandwich born in East Harlem and the South Bronx, made in bodegas usually owned by Yemeni immigrants, with cooks who are often Dominican or Mexican, served to a clientele who are mostly Black and Latino, and which until about 2016 didn’t have to share its existence with the rest of the country. (Then it went viral and there were fifteen-dollar versions in Manhattan and a real argument about cultural appropriation, which is its own essay.) It is also the New York whose lexicon has been rapidly going national: deadass, mad as an intensifier, brick for cold, son and b as vocatives, much of the glossary of drill rap. If you’ve heard a teenager in Ohio say deadass, that teenager is talking like the Bronx. The chopped-cheese New York is the engine of the city’s linguistic export economy.

The BEC New York (the Bodega Cosmopolitan)

The bacon-egg-and-cheese is the cross-cutting case. It’s everywhere because the bodega is everywhere. The comedian Desus Nice has a much-quoted bit about what makes a real bodega — bulletproof glass, a cat, a guy at the counter named something like Mohammed or Papi or Ahki — which sounds like a joke and is actually a Hymes-style ethnography of a speech event compressed into a single sentence. Setting: the bodega. Participants: you and a guy whose first language is probably not English but whose ordering vocabulary is absolutely fluent. Act sequence: “BEC SPK on a roll.” Norms: don’t dawdle, exact change appreciated, “thank you boss” exits the conversation. The fact that baconeggandcheese has genuinely lexicalized into one word — not three — is itself a linguistic feature of the variety. You don’t say bacon, egg, and cheese. You say one thing.

You speak BEC at the bodega; you go home and speak something else.

The BEC New York is what you get when all three speech communities have to share a counter. It’s the closest thing the city has to a true contact zone — a place where features pool. But it’s a transactional community, not a residential one. You speak BEC at the bodega; you go home and speak something else.

Why the European Theory Doesn’t Quite Work Here

The reason this matters — the reason it’s interesting and not just a list — is that European sociolinguists have spent the last twenty years studying cities that look superficially like ours, and they have mostly arrived at one finding: a multiethnolect. Multicultural London English is the canonical case. Jenny Cheshire, Paul Kerswill, and their collaborators showed that in inner-city London, second-generation kids from Caribbean, South Asian, African, and White working-class backgrounds had collectively built one shared variety: monophthongal vowels, man used as a generic pronoun, a new quotative this is + speaker. Heike Wiese has documented Kiezdeutsch in Berlin among Turkish-, Arab-, and German-heritage youth; Pia Quist has documented similar varieties in Copenhagen; Chege Githiora has written a whole book about Kenya’s Sheng, which fuses Swahili and English and serves as a detribalizing language for young Nairobians. The general pattern is that when a major city has heavy youth contact across many heritage backgrounds, you get one consolidated variety.

But there’s a catch, and it’s a big one: When Cheshire’s team applied the same research design in Paris — the same migration profile, the same age cohort, the same data collection — they didn’t find an MLE-equivalent. No consolidated multiethnolect emerged. The conditions were different in some way they’re still working out, but the takeaway is clear: multiethnolect emergence isn’t automatic. It depends on the geography of contact.

That’s where New York gets interesting. Our immigration didn’t concentrate in one inner city. It dispersed across five boroughs and dozens of ethnically marked neighborhoods that, until the subway brings them together at a public school, mostly keep their own counsel. Caribbean Brooklyn and Dominican Washington Heights and Mexican Bushwick and South Asian Jackson Heights and Bukharian Forest Hills are all New York, but they aren’t all the same New York — and they don’t necessarily produce one consolidated youth variety the way inner London did. They produce parallel varieties. The chopped-cheese vernacular is one of them; it happens to be the one that has gone viral. There are others we know almost nothing about.

The Queens Gap

The sandwich frame has gotten me this far, but it has a built-in limit I should name. The three iconic sandwiches I’ve been using as my diagnostic — bagel-and-lox, chopped cheese, BEC — all come from waves of New York immigration that have already been culturally canonized. They’ve had decades of journalism, film, food writing, and folklore to enshrine them as the sandwiches. The newer waves of immigration haven’t yet produced sandwiches that show up on anyone’s top-three list, and the food writers haven’t caught up. That’s actually the point. The gap in the sandwich canon mirrors a gap in the linguistic literature. The communities whose foods haven’t been canonized are the same communities whose speech hasn’t been studied. Both gaps come from the same place: where attention has decided to look.

I want to spend the rest of this essay on Queens specifically, and I should say up front why. It’s because patchworks like the one I’m about to describe exist in every borough, and someone could reasonably ask why the borough I happen to be from gets the spotlight. Two reasons. The first is that Queens is, by most measures, the most linguistically diverse county on Earth — somewhere north of 160 languages spoken in everyday use, the highest density of language diversity anywhere we know of. The Endangered Language Alliance is based in New York largely because of what’s happening in Queens. So if you want a single test case for the parallel-speech-communities argument, Queens is where the parallels are stacked deepest.

The second reason is that Queens lacks a dominant ethnic frame in a way the other boroughs don’t. Brooklyn has its Jewish-Italian-Caribbean narrative; the Bronx has its Puerto Rican-and-AAVE-and-hip-hop narrative; Manhattan has its Lower East Side and Harlem narratives. Each of those frames organizes how researchers and journalists approach the borough. Queens doesn’t have a singular frame because no one community has ever dominated — which makes it both the clearest case for parallel speech communities and, conveniently for my argument, the least studied. The other boroughs have their own writers and their own future field researchers. I’m just the one who grew up in Queens.

So I’m a child of Queens in a way that I’d argue qualifies me — with the appropriate humility — to make claims about its linguistic terrain. I’ve lived in Jamaica, and have been schooled in Forest Hills, Rego Park, and Flushing. I worked in others. And I made a deliberate practice, growing up, of not just letting contact with other communities happen organically but of seeking it out, spending afternoons in ethnic enclaves I had no family ties to, eating at restaurants where I was the only person not speaking the kitchen’s language, going to community events for traditions that weren’t mine. By the time I got to college I had hung out, eaten, prayed near, studied with, and gossiped alongside people from most of the borough’s significant immigrant communities. My chest swells with a small amount of pride about this. I’ve seen a lot.

One concrete consequence of all that contact: I’m equally at home with both sides of the linguistic divide this essay opened with. I’m as comfortable with a well-deployed schlep or kvetch as I am with mad as an intensifier or deadass at the end of a sentence — though you’d be hard-pressed to catch me using either variety out loud without at least a flicker of self-awareness, since neither variety is mine by birth. The point is the receptive fluency, not the performance. I know what these words mean, when they’re being used straight, when they’re being used ironically, what each one signals about the speaker, and which combinations sound natural and which sound off. That came from years of being in rooms where people spoke this way, on both sides, and paying attention.

And yet, as a budding linguist, I have to concede immediately that I’ve barely scratched the surface. Lived experience and ethnographic observation aren’t the same thing, and ethnographic observation itself isn’t the same as variationist data. The thing I can say from inside the borough is that there is much more here than the published literature reflects.

The streets of Forest Hills and Rego Park are trilingual at minimum — English, Russian, and Bukhori within a few blocks of each other; Punjabi, Urdu, and Hindi a few stops east; Bengali and Nepali a few stops further. I’ve been studying Russian for a while now, partly out of academic interest and partly because the Bukharian Jewish community of Forest Hills and Rego Park is one of the most linguistically rich pockets in the city — tens of thousands of speakers operating across English, Russian, and Bukhori (a Judeo-Tajik variety), often within a single conversation. To even hear what’s happening in that community linguistically you need at least some Russian, and the published sociolinguistic literature on it is essentially nonexistent (or at least invisible to an amateur like me). The Indo-Caribbean community in Richmond Hill and South Ozone Park (eighty thousand foreign-born Guyanese alone, plus Trinidadians and Surinamese) sits at the intersection of Caribbean Creole, North Indian heritage, AAVE-influenced peer English, and the new chopped-cheese vernacular all at once. The South Asian neighborhoods I grew up in — Punjabi Sikhs running gas stations on Atlantic Avenue, Bangladeshi families in Jamaica, Gujarati shopkeepers in Jackson Heights — are similarly invisible to the literature.

We have demographic data; we have rich journalism; we don’t have variationist studies.

Almost none of this has been documented sociolinguistically. Devyani Sharma has done serious variationist work on second-generation Punjabis in London’s Southall — generational style repertoires, selective retention of substrate features, the works — and that’s the obvious comparative model. But nobody to my knowledge has done the equivalent for Jackson Heights or Richmond Hill or Forest Hills. We have demographic data; we have rich journalism; we don’t have variationist studies. This is, I suspect, where the most interesting research on New York English is going to come from in the next decade. Whether the Queens varieties will eventually feed back into the chopped-cheese vernacular and produce something more like an MLE-style consolidation, or whether they’ll stay parallel — that’s a real empirical question, and it hasn’t been answered.

The Sandwich, Returned

Back to lunch. The reason we have three sandwiches that everyone agrees are New York is the same reason we have three speech communities that don’t quite see each other — and a Queens patchwork that few have yet looked at. Each sandwich tracks a different cultural formation, with its own history, its own ordering ritual, its own lexicon, its own Hymes-style speech event. The bagel-and-lox crowd doesn’t know what a chopped cheese is, and the chopped-cheese crowd thinks lox is something that goes on a salad bar. Both of them order BECs at the same bodega. None of them, mostly, order roti and dal at the Trinidadian shop on Liberty Avenue, even though it’s all in the same city.

You sound like your sandwich. Or, more carefully: you sound like the speech community whose sandwich you grew up ordering. New York is not one accent dying. It’s three or four accents talking past each other on the F train, plus a fifth I suspect we’re going to find as soon as somebody goes looking.