A talk is not a paper

Claim: a talk is its own genre: it says its point and plan early, returns to them, and ends on the point.

What it asks

  • State the one thing to remember in the first minutes (Peyton Jones 1993; TED 2026, example only).
  • Show the plan, as an organiser. The rule is overview first.
  • Return to the point at the end of each section, and end on it.
  • Split long material into segments (Rey 2019).
  • Do not read the paper aloud: its structure is not the talk's.

Why

In 44 talks in geology, medicine and physics, introductions were structured markedly differently from the paper's. The authors tie this to the contextual and epistemological features of the genre (Rowley-Jolivet 2005).

Segmenting gave small to medium gains in retention and transfer across 56 studies and 88 comparisons (Rey 2019, abstract).

Limits

  • The genre research describes what speakers do. It does not show that the structure helps a listener.
  • No review of talk structure was found. Carter-Thomas 2003 is a methods paper; Hood 2005 and Dubois 1980 were found by title, not read.
  • Peyton Jones 1993 advises skipping the outline slide, and placing a recap slide about a third of the way through. That is advice from experience, untested; it disagrees with a plan slide at the start, and agrees with returning to the point.
  • Ending on the point is a convention. A serial-position effect on recall might favour it, but no source for this was kept.
  • The belief that attention drops after 10 to 15 minutes has little primary support (Wilson 2007; Bradbury 2016), so it does not justify the repeats.
  • Experts agree on talk advice, but experiments are scarce (Blome 2017, 91 expert-opinion papers).