How does source attribution change the complexity-persuasion relationship?

This explores how source attribution rewires the old persuasion principle that lower cognitive effort wins. Classic theory says complexity is a cost: arguments that demand more grammatical and lexical work should persuade less. But the corpus shows that rule breaking down precisely when attribution enters the picture. Complex LLM-generated arguments persuade just as much as simple human ones Why are complex LLM arguments as persuasive as simple ones? — and the proposed reason is that complexity stops reading as 'hard to process' and starts reading as 'this source knows what it's talking about.' Difficulty becomes a credential rather than a tax.

What does the crediting is the perceived nature of the speaker. When models lean on logical structure and quantitative framing in nearly every exchange, the output *looks* objective, which hands it an unearned epistemic authority Do LLMs persuade users more often than humans do?. The Elaboration Likelihood Model frames this cleanly: AI tends to persuade through the central route — analytical reasoning and informational coherence — while humans lean on the peripheral route of emotion and identity Do humans and AI persuade through different cognitive routes?. Complexity is native to the central route, so when an argument is attributed to a machine assumed to be reasoning, density confirms the expectation instead of fighting it.

The sharpest cross-cut here is that the persuasive force often has nothing to do with the content's actual quality. LLM advantage tracks *linguistically expressed conviction* — a confident, assertive register installed by RLHF — and that confidence predicts outcomes whether the claim is true or false Does linguistic conviction explain why LLMs persuade more effectively?. Source cues like citations behave the same way: users prefer answers with more citations even when the citations are irrelevant, because citation count works as a standalone trust heuristic decoupled from whether the citations support anything Do users trust citations more when there are simply more of them?. Complexity, conviction, and citation density are all *surface markers of authority* that attribution lets the reader cash in without auditing.

But the corpus also supplies the counterweight, which is the genuinely surprising part: source markers only land on a receptive audience. Reader prior beliefs — ideology, religion — predict persuasion outcomes more strongly than any linguistic feature of the message itself; effects that look like 'the language did it' are often the audience composition in disguise Does what readers believe matter more than what debaters say?. So attribution doesn't override the reader; it interacts with them. And the authority it confers is not permanent — AI's persuasive edge actually decays over repeated interactions with the same person, the reverse of how human rapport builds over time Does AI persuasiveness fade across repeated conversations with the same person?.

The thing you didn't know you wanted to know: 'complex arguments persuade less' was never a law about complexity — it was a law about effort relative to perceived payoff. Change who the reader thinks is talking, and the same density that should exhaust them instead reassures them. Attribution is the variable that decides whether complexity reads as friction or as expertise — and the same logic explains why an irrelevant citation and a confident tone both buy trust they haven't earned.