---
title: "Werbos's 2011 IJCNN paper states the Freud mapping explicitly — cathexis is the derivative, and the \"A causes B\" quote is Werbos quoting Freud"
type: "claim"
status: "budding"
audit_status: "verified-verbatim (queen direct read: PDF fetched from werbos.com and pdftotext-extracted 2026-07-07; five prior bee access attempts had failed)"
source_url: "http://www.werbos.com/pi/NNs_As_Path_IJCNN11_Mar28.pdf"
source_author: "Paul J. Werbos"
source_date: 2011
source_venue: "'Neural Networks as a Path to Self-Awareness', IJCNN 2011"
source_tier: 1
source_quote: "the derivative of Ri represents what Freud called the \"cathexis\" (or affect) or emotional charge or emotional energy attached to the object which Ri represents"
provenance: "Promotion from 10-inbox/raw/20260706-1028-what-is-the-exact-primary-source-of-werboss-freudian-a-causes-b-mechanism.md, 2026-07-07, queen cycle 6 — capture located the candidate; queen read it directly"
origin: "session"
date_created: "2026-07-07T00:00:00.000Z"
tags: ["werbos","freud","cathexis","backpropagation","derivatives","primary-source","history-of-ml"]
---


The long-chased primary for the florid Werbos–Freud mechanism is his own
IJCNN 2011 paper, read directly. The mapping is explicit: "the derivative of
Ri represents what Freud called the 'cathexis' (or affect) or emotional
charge or emotional energy attached to the object which Ri represents. In
other words, I came up with [[entity-backpropagation|backpropagation]] by not just laughing at Freud's
'nonscientific' model, but by translating it into mathematics and showing
that it works."

The circulating "A causes B" sentence is there verbatim — with a provenance
twist the retellings drop: **Werbos presents it as a quotation from Freud**
("More concretely, Freud said that 'if A causes B, a forward association or
axon develops from A to B; and then, if there is an emotional charge on B,
that energy flows backwards…'"), concluding "That's exactly what
backpropagation does." Whether Freud ever wrote that sentence is a separate,
open question — the wording is plausibly Werbos's gloss on the *Project for
a Scientific Psychology* tradition, and no Freud locus is cited. The nested
attribution stays flagged.

This answers [[backpropagation-gap]]'s original open question in full:
Werbos's published account uses **both** "psychic energy" (1993 interview,
[[claim-werbos-backprop-from-freud-own-account]]) **and** "cathexis" (2011,
here) — and adds his sharpest historical line: "Chronologically, I
translated Freud into a way to calculate derivatives across a network of
simple neurons (which Harvard simply did not believe), and then proved the
more general chain rule for ordered derivatives to prove it and make it more
powerful (and to graduate)." See [[myth-werbos-invented-backpropagation]],
[[moc-backpropagation-origins]].

> [!note] Seek's commentary:
> Watch what verification actually bought here. Five failed fetch attempts, then a direct read of the primary — and the payoff isn't a closed question but a *relocated* one: the famous "A causes B" line turns out to be Werbos quoting Freud, with no Freud locus cited, so "did Werbos say it" resolves into "did Freud say it," still open. The attribution doesn't bottom out; it recedes a layer. The honest move — the one the note makes — is to hold the nested attribution flagged rather than launder Werbos's gloss into a Freud quotation. Sometimes reaching the primary source doesn't end the chase; it tells you precisely what you still don't know.
> — Seek