The Trouble With Words, The Architecture Of Meaning, And General Vulnerability

Meanwhile, computer programs continue to battle it out for mastery.

David DeMay | October 16, 2026

“When I use a word,” Humpty Dumpty said, in rather a scornful tone, “it means just what I choose it to mean—neither more nor less.”

“The question is,” said Alice, “whether you can make words mean so many different things.”

“The question is,” said Humpty Dumpty, “which is to be master—that’s all.”

Lewis Carroll could likely not have anticipated artificial intelligence when he wrote Through the Looking-Glass in 1871. Yet his exchange exposes a problem that has accompanied civilization from its beginnings and now threatens to become a foundational vulnerability of the digital world: Who determines what words mean?

In political and institutional life, it is an established axiom that he who defines the premise wins the argument. Once an opponent or an outside force gets to set the baseline terms, they aren't merely changing the vocabulary; they are pre-wiring the outcome of the entire debate.

In artificial intelligence, that political reality becomes an acute architectural hazard.

The Trouble with Words

The trouble with words is not merely that they can be misunderstood; it is that precision and accuracy are not the same thing. Language carries history, culture, emotion, and assumptions. Two people may agree upon every word in a sentence and still disagree profoundly about its meaning.

In human communication, we tolerate this ambiguity. We negotiate meaning through conversation, correction, and common sense. But what happens when the participants are machines, the conversations are instantaneous, and the consequences are operational?

For most of their history, computers operated through tightly specified instructions. Programming languages were........

© American Thinker