Dr. Nabil is sixty-eight. He is my uncle’s best friend, though you will not know him and that is not his real name. He smokes too many Marlboros, drinks his coffee bitter, no sugar, a tiny cup lasting half an hour, and still publishes essays on Arabic poetry that make you feel the language is a living body, not a set of rules.
His wife Layla notices things. A pair of glasses left inside the fridge, a name that arrives a breath too late. She brings him to a neurologist on Clemenceau Street, a good man trained in Paris. The doctor administers the test in French, Nabil’s intellectual language. Nabil names the camel, the rhinoceros, the date. He draws the clock face, hands at ten past eleven, and the circle is perfect. His recall of five words is flawless. Twenty-nine out of thirty. The neurologist smiles, pats his shoulder, and tells Layla it is normal ageing, a little slowing, nothing to worry about.
Two years later Nabil cannot follow a single sentence in a conversation. The pathology had been chewing through his entorhinal cortex all along, but the test had been listening for the wrong thing, in the wrong language, in the wrong cognitive world.
I want to tell you why this happens, and why it happens so often in Lebanon that it has become a silent, private catastrophe, repeated in thousands of living rooms from Tripoli to Nabatieh. The problem is not a shortage of clever neurologists or a lack of scanners. The problem is that the cognitive tests we use were born in laboratories in Boston, London, and Paris, and they were designed for monolingual brains. They ask a patient to name a static object, to recall a list of words, to copy a shape. They do not ask the one thing a Lebanese brain has been doing every single day since childhood: moving between languages, selecting one while suppressing another, switching mid-sentence, mid-thought, mid-breath. And that exact ability, the constant, invisible, beautifully Lebanese juggling of tongues, is among the first neural faculties to fray when Alzheimer’s begins its quiet work.
The particular problem this essay is about is the diagnostic silence that falls over multilingual, diglossic populations when they are tested with monolingual tools. A Lebanese elder who speaks colloquial Arabic at home, Modern Standard Arabic in formal settings, and French or English in professional life has a brain that is structurally different from a monolingual brain. Lifelong code-switching remodels the anterior cingulate cortex, thickens the left caudate, and reinforces the white-matter tracts that connect the prefrontal cortex to the deeper language regions. This neural architecture is a tremendous cognitive gift. It is why bilinguals, on average, show the clinical symptoms of Alzheimer’s three, four, even five years later than monolinguals with the same biological burden of amyloid plaques and tau tangles. The brain compensates. It borrows from the overbuilt executive networks. And for years, the compensation works so well that a standard cognitive screen sees nothing at all.
But the gift masks the disease.
A monolingual English speaker with mild entorhinal tau accumulation will struggle on a category fluency task: “Tell me all the animals you can think of in one minute.” The retrieval network is already compromised, and the word count drops. The test catches it. The Lebanese bilingual facing the same pathology, asked the same question in French, will glide over the gap. When the French word for “giraffe” stalls, the Arabic word “zarafeh” arrives automatically, unconsciously, and the total word count stays normal. The clinician records “cognitively intact” and the pathology advances another two years into brain regions from which recovery is far less likely. By the time the system collapses and the family finally receives a diagnosis, the patient is already in moderate dementia, and the window for early intervention has closed.
This is not a hypothetical concern. Lebanon has one of the fastest-ageing populations in the Arab world, the highest dementia risk factors, diabetes, hypertension, the chemical residue of decades of stress, and an almost perfect storm of linguistic multiplicity. We do not merely speak two languages. We live inside them. A morning in Beirut may open with “bonjour, kifak” and move through a radio bulletin in Modern Standard Arabic, a business call in English, an argument about politics that fuses all three in one sentence, and an evening with a grandmother who understands the television news in Fusha and answers you back in the Amiyah of her mountain village. Every single one of these transitions requires the brain to inhibit one linguistic system and activate another, to monitor for errors, to select the right word from a competing set while the previous language is still humming just beneath the surface. This is not a parlor trick. It is an executive workout of formidable intensity. And it depends on the precise frontal and striatal circuitry that Alzheimer’s pathology invades early.
The scientific bottleneck, then, is piercingly clear. We lack a brief, culturally truthful, language-switching stress test, a way of deliberately taxing the bilingual executive system to see if it is beginning to fail. Standard tools never ask the brain to switch. They let it stay comfortably inside a single language, where automatic compensation can mask the cracks. The solution this essay proposes is a diagnostic concept I will call Lisan. In Arabic the word means both “tongue” and “language,” and the test is built on a very simple insight: if early Alzheimer’s whispers first in the space between a word and its translation, we should build a tool that listens exactly there.
Imagine a quiet room in a clinic in Ras Beirut. A laptop on a small table. A patient, perhaps a retired literature professor like Dr. Nabil, perhaps an illiterate grandmother from the Beqaa who has never held a pen, sits comfortably and looks at the screen. A picture appears: a cup of coffee, a cedar tree, an olive, a mosque. Just before the picture, a small flag flashes, the Lebanese cedar for colloquial Arabic, the classical Tawhid flag for Modern Standard Arabic, the French tricolour for French. The patient’s only task is to name the picture aloud in the language the flag demands. On some trials the flag repeats the previous one; on others it switches. The whole session takes twelve minutes.
The science lives in the pause. A healthy multilingual brain pays a small, predictable cost to switch, perhaps fifty or sixty milliseconds of extra processing time, a barely perceptible hesitation. But a brain in which the fronto-striatal switching circuitry has begun to accumulate tau pathology pays a larger and larger cost. The hesitations stretch. Errors creep in. The person might name the coffee cup in the wrong language and correct themselves a moment too late. And this happens while their static naming still looks perfect to a standard test. The switch cost becomes the true biomarker, a signal that does not depend on education, or literacy, or which languages you happen to know, because it is measuring the underlying neural operation, not the surface vocabulary.
The beautiful thing about Lisan, the genuinely Lebanese thing about it, is that it works equally for everyone because it uses the diglossia that already lives in every Arabic speaker. Lebanon is not just bilingual; it is diglossic. Colloquial Lebanese Arabic and Modern Standard Arabic are highly distinct linguistic systems, with different phonology, different syntax, different vocabulary. Every Lebanese person, literate or not, navigates between them daily. The grandmother who cannot read a newspaper understands the television news in Fusha and answers her daughter in Amiyeh. Her brain performs the same executive shift as the trilingual professor’s. So the test can be offered in two parallel, neurologically equivalent forms: a Colloquial–MSA version for elders with little formal schooling, and an Arabic–French version for the highly educated. The brain circuit under interrogation is identical. The test is schooling-blind, not because it ignores education, but because it listens to a mechanism that education does not build and illiteracy does not weaken.
How would a small team begin turning this idea into reality? Not with an engineering lab or a million-dollar grant. With a few months of careful, human-scale groundwork. A researcher, a clinical collaborator, a research assistant who speaks the dialects fluently, this is enough. First, you gather sixty culturally rooted images: a cup of coffee, an argileh, a sprig of za’atar, the Raouche rocks, a plate of tabbouleh. You show them to thirty healthy young Lebanese adults to make sure everyone agrees what each picture is called in each language. You pilot the switching task with fifteen healthy elders to see that it feels natural, that the flags are understood, that nobody is anxious or confused. You adjust. You laugh with them when they argue about whether the picture shows oregano or za’atar. This pilot phase is not just methodology; it is the work of making a test that feels like Lebanon, not like a laboratory.
Then the real study. Ninety participants, aged sixty to eighty-five, recruited through the geriatric and neurology clinics of a Beirut hospital, AUBMC, Hôtel-Dieu, maybe the Hariri public hospital so the sample is not only private and wealthy. An independent neurologist diagnoses them into three groups: thirty with amnestic mild cognitive impairment, thirty with mild probable Alzheimer’s, and thirty healthy controls. Each participant completes the Lisan switching task, a simple history of which languages they use and how often, and maybe one other quick measure of non-verbal inhibition, just to confirm that any deficit is not purely linguistic but reflects a broader executive decline. Everything is recorded on a cheap USB microphone. Reaction times are extracted later from the audio files using free software. Accuracy is scored by a native speaker listening carefully, noting not only whether the word was right but whether there was a false start, a self-correction, a slip into the wrong language that was caught too late.
The analysis asks a clean question. Does the switching cost, the extra time and the extra errors on switch trials compared to non-switch trials, discriminate between the healthy group and the earliest clinical stage of Alzheimer’s? And does it do so with equal strength in the high-education, French-speaking group and the low-education, Colloquial–MSA group? The prediction is that it will. That the switch cost will detect amnestic mild cognitive impairment with significantly higher sensitivity than the standard cognitive screen, especially in the high-reserve, highly educated patients whom the current tools systematically miss. That a twelve-minute test built on a laptop and a set of culture-true pictures can hear what a full neurological consultation cannot.
The deliverable at the end is not a product to be sold. It is a fully standardised, open-source protocol and a set of validated stimulus images, freely available for any clinician in any corner of the Arabic-speaking world to download and use. A test that costs almost nothing to administer and that finally treats multilingualism not as a confounding variable that makes testing harder, but as the diagnostic instrument itself.
I keep thinking about Dr. Nabil. About the two years between that perfect score and the afternoon he could no longer follow a sentence. Those two years were not empty. His family made decisions in the dark, attributed his small slips to stubbornness or ageing or the stress of a country that has been breaking for as long as anyone can remember. They needed a test that could have looked at him, at his magnificent, multilingual, poetry-filled brain, and said something is beginning to change. Not in the words he still knew perfectly, but in the tiny, costly pause between them.
Lisan is an attempt to build that test. To press the stethoscope where the pathology sings loudest. To take the most Lebanese thing about us, our movement between worlds of language, and make it the thing that saves us, finally, from being heard too late.