Narrative as Evidence

In the AI Story

Aristotle's original argument supplies the philosophical foundation: the phronimos learns not from rules but from cases. The cases do not dictate the right answer for future situations, because future situations differ in ways no rule can anticipate. The cases provide something more valuable than answers: material for the development of judgment. The reader who engages with a well-rendered case develops, through that engagement, a capacity for reading situations that no abstract principle can substitute.

The most revealing passage in Segal's Orange Pill, from a phronetic perspective, is not a triumph but a failure: the author's description of catching Claude attributing a concept to Deleuze that the philosopher never articulated, a passage rhetorically elegant and philosophically wrong. The aggregate finding — LLMs produce factual errors at rate X — is episteme. It tells you what happens on average. It does not tell you what it feels like to almost accept an elegant falsehood because the falsehood was smoother than anything you could have produced yourself, or how to develop the specific kind of attention that detects the seam where confident wrongness meets good prose. The narrative tells you this.

The reader of phronetic narrative engages differently than the reader of quantitative research. The quantitative finding is received passively — it holds or does not, regardless of reader judgment. The narrative is received actively — the reader must decide what the account means for her circumstances, what lessons to draw, what cautions to heed. The reader's phronesis is not optional. It is constitutive of the knowledge the narrative produces. The same narrative, read by different practitioners in different contexts, generates different phronetic insights, because the insights are produced in the encounter between narrative and reader's judgment, not in the narrative alone.

The institutional exclusion of narrative evidence is not accidental. Academic journals favor quantitative studies. Funding agencies reward replicable designs. Tenure committees count citations, which cluster around abstract findings other researchers can build upon. The incentive structure systematically disadvantages the form of research the AI transition most urgently requires. Flyvbjerg has argued throughout his career that the case study is not a preliminary stage in the production of real knowledge but a form of knowledge in its own right — the form best suited to context-dependent phenomena. The AI transition has made this argument operationally urgent, because the evidence base being built on quantitative instruments systematically misses the phronetic dimension that matters most.

Origin

The methodological argument is developed across Flyvbjerg's work, most fully in Making Social Science Matter (2001) and 'Five Misunderstandings About Case Study Research' (Qualitative Inquiry, 2006). Its application to AI narrative evidence is developed in the 2025–2026 work.

Key Ideas

Terminal, not preliminary. Narrative is not a first stage awaiting statistical aggregation; it is the form of evidence adequate to context-dependent phenomena.

Judgment-building. Cases develop the reader's capacity for situated judgment in ways that abstract principles cannot.

Active reception. The reader's phronesis is constitutive of the knowledge narrative produces — the same case generates different insights for different readers in different contexts.

Institutionally excluded. Academic incentive structures systematically disadvantage the form of research the AI transition most needs.

Complement, not replacement. Narrative and quantitative evidence illuminate different dimensions of the same phenomenon; the argument is for inclusion, not substitution.