A second, directly-read 1988 Roberts interview (Pelkey/CHM) independently corroborates the Mansfield-era DARPA pressure pattern, without Werbos's 'crisis warning' phrase

larry-robertspaul-werbosdarpamansfield-amendmentchm-pelkey-oral-historyscience-policy

In a June 1988 Computer History Museum oral-history interview conducted by James L. Pelkey, Larry Roberts discusses the 1969 Mansfield Amendment in his own voice, on a document distinct from both the Bartee 1988 book chapter and the 2025 Gilliam essay that the vault's existing Roberts/Werbos corroboration relies on. Asked by Pelkey about the "critical decision... forcing where you spent your money, to be much more focused on immediate defense applicable projects, as opposed to more fundamental research," Roberts replies: "It was, and we had to keep on working to get around it, but it got harder and harder with time." In the preceding lines Roberts dates the pressure himself — Pelkey: "When Mansfield came in in the early '70s --"; Roberts: "Yeah, probably '71" and, asked whether he was still at ARPA, "I had to get around it."

Two provenance caveats the transcript itself carries, added by the 2026-10-19 audit: the exchange sits in the second of the two Pelkey interviews, whose exact date is unrecorded (the editor's note dates it only "a few days after" the June 17, 1988 first interview), and it falls on pp. 32–33, inside the span the transcriber flags from the final tape-side break (p. 26) onward: "The following material is of marginal reliability at best, due to extremely poor quality of the recording." The quoted lines are as the transcript prints them; the transcription layer under them is flagged by its own transcriber.

This is the same witness on a second, independent record (a different interviewer, a different sitting, read directly rather than through a secondary essay) of the same underlying pattern: Mansfield-era pressure forced IPTO away from open-ended research toward defense-applicable output, strong enough that Roberts says he "had to get around it." It corroborates the mechanism Werbos describes for his own DARPA program — a forced reorientation away from "pure theory" — without using Werbos's specific phrase "crisis warning," which is tied to Werbos's own forecasting-adjacent work rather than Roberts's networking portfolio. The phrase itself remains particular to Werbos's testimony; what transfers is the shape of the pressure, not the vocabulary.

This note does not resolve whether the Gilliam-quoted fragments in the existing corroboration note actually sit inside Roberts's own 1988 Bartee chapter (see question-does-roberts-1988-bartee-chapter-or-cbi-oh159-confirm-gilliam-quoted-fragments) — it is a separate primary, not a check on that one.