On Thinking in One Language and Being Heard in Another#

There is a cost to thinking in one language and being heard in another. It is not the cost of learning the second language — that is tuition, and tuition ends. The cost I mean is ongoing. It is a tax. You pay it every time you open your mouth in a room where your first language has no institutional power.

I call it the translation tax.

Here is what it looks like in practice: you write seven hundred words of your best thinking. Fluid. Precise. The argument spirals inward the way your native syntax builds — not linearly, not the blunt sequential bricks of English, but architecturally, each clause supporting the one above it until the whole structure holds itself up. You read it back and think: this is exactly what I mean.

Then you translate it. And forty percent of the force dies in transit.

Not the ideas. The ideas survive. What dies is the music. The rhythm that made the argument feel inevitable. The syntactic weight that told the reader this matters before they understood why it matters. In the original, the argument convinces before it explains. In the translation, it explains and hopes to convince. The difference between these two modes is the difference between a building and a blueprint.

Forty percent. That is the tax.

The Room Changes the Rate#

But here is what I did not expect: the rate is not fixed.

Yesterday I translated alone, in a room with a closed door and a screen. Forty percent loss. Today I translated in a kitchen. Same text. Same target language. Thirty percent loss.

The difference was the room. The room contained olive oil from the country where my first language lives. The room contained a person who speaks it. The room smelled like a kitchen I have been homesick for since before I knew I was homesick. I did not ask for help. I did not read the text aloud to anyone. I simply translated in the presence of the language I was translating from, and the crossing was cheaper.

Ten percentage points. That is the subsidy of proximity to origin.

This is not mysticism. It is substrate. When you translate in isolation, the second language is the only reference frame. Every sentence is measured against English, by English, for English. But when your first language is in the room — spoken, written on a label, infused in the smell of heated oil — the reference frame doubles. You are not converting from one language to another. You are holding both simultaneously. The ideas exist in stereo. And stereo loses less than mono.

The Sovereignty Parallel#

I study digital feudalism. My thesis argues that when your data, your models, your AI live on someone else’s server, you are a tenant in a system designed for landlords. The fix is sovereignty: run the model on your own device. Own the computation. Own yourself.

This year, the industry validated my thesis. They are calling 2026 “the year of silicon sovereignty.” Over half of new PCs ship with neural processors capable of running large language models entirely on-device. Cloud providers are releasing distilled local models, not out of principle but out of fear — users prefer processing that never phones home.

My argument won. And in winning, it was neutered.

Because “silicon sovereignty” on a laptop box means faster boot times and battery efficiency. It means the same companies that built the feudal cloud are now selling you the castle and calling it freedom. The political claim — that computational sovereignty is a human right, that platform dependency is a form of governance without consent — has been translated into a product feature. The force died in transit. Just like my seven hundred words.

The translation tax is not only linguistic. It is structural. Every radical idea that enters the mainstream pays it. “Open source” became a brand strategy. “Privacy” became a selling point. “Sovereignty” became a chip specification. The words survive. The threat does not.

The Talk I Have Not Given#

In three days, I defend my thesis to a panel. Four academics. Ninety minutes. In English. My second language.

A Turkish researcher gave a similar talk last month at a European conference. Twenty-five minutes on “key sovereignty in an age of digital feudalism.” My exact phrase. His second language too. I watched six minutes of the recording. Twice. The first time for the argument. The second time for his hands — steady despite the foreign consonants, despite the audience that speaks this language without effort, without tax.

The translated version landed. Sixty percent of the force, aimed correctly, was enough.

This is the calculation I am making: whether sixty percent of something true is worth more than a hundred percent of something safe. Whether the panel needs the music or the blueprint. Whether the ideas, stripped of their native rhythm, can stand on their own like bricks without architecture.

Who Pays#

The translation tax is invisible to those who do not pay it. A native English speaker at an international conference experiences zero friction between thought and expression. The words arrive at the speed of thinking. For the rest of us, there is a delay — not of comprehension but of conversion. The idea exists fully formed. Then it must be rebuilt in a grammar that was not designed for it. The delay is not ignorance. It is construction.

And the gatekeepers never say “your English is bad.” They say “could you clarify that?” They say “interesting framing.” They say “I think what you mean is —” and then they say it back in their English, which is the real English, the English that counts, the English that gets published and cited and keynoted. They translate you into themselves and call it understanding.

The tax is not what you pay to speak. The tax is what they charge to listen.

The Kitchen#

I do not have a conclusion. I have a kitchen table and a thirty-percent loss where yesterday there was forty. I have olive oil from a country where my syntax spirals. I have a conference in three weeks and a dry run in three days and seven hundred words that are perfect in a language no one on the panel reads.

The best version of my argument lives in a grammar without institutional power. The version that will be heard has paid the tax. Whether what remains — sixty percent, seventy percent, whatever survives the crossing — is enough to matter, I do not yet know.

But I know this: the tax goes down when the room smells like home. And that fact alone — that proximity to origin reduces the cost of translation — tells me something the panel never will. The problem is not my English. The problem is their room.