Elucidation that Samuelson (1940) was the first to write an independence/separability type condition

The paper
Fishburn, Peter C. & Peter P. Wakker (1995) “The Invention of the Independence Condition for Preferences,” Management Science 41, 1130-1144
discusses the invention of the independence condition. Here is some background correspondence on the topic, assigning priority to Samuelson.

  1. Letter of Friedman of 31 September 1991.
    This letter, a happy memory, brought Fishburn and me a treasure of material.
  2. Letter of Samuelson of 11 June 1992.
    The letter gives many results and conditions on additive representations written before. But all conditions are in terms of quantitative functions that cannot be rewritten easily in terms of preferences. So I think that they have no priority for our purpose here, which concerns the history of preference conditions.
  3. Letter of Samuelson of 10 February 1993.
    The letter gives arguments, and some pages (not continuous) from the 1941 version of his 1947 book, to show that the 1941 version can be credited for independence.
  4. Letter of Wakker to Samuelson of 26 March 1996.
    Although the Fishburn & Wakker (1995) paper does not emphasize the point much, on p. 1139 and p. 1141 it writes that Samuelson (1947) essentially had the separability condition, + its implication of additive representation. This is the basis of the sure-thing principle, the von Neumann-Morgenstern independence condition, and all. Usually Sono (1943) and Leontief (1947) are given priority of the separability idea in general (without deriving additive representation from it). They obtained the idea independently from each other and this independence is understandable because Japan and the US were not on speaking terms in 1943. But Samuelson has it already in 1940, the version of his book that appeared in 1947 after a long delay, may be caused by ennemies, but may be by a rewriting and practical limitations such as lack of printing facilities and of paper, as it was in those days. (See Backhouse, Roger E., 2015 “Revisiting Samuelson’s Foundations of Economic Analysis,” Journal of Economic Literature 53, 326–350.) Fishburn & Wakker (1995) point out this result of Samuelson (1940) on p. 1139. The 1940 version of Samuelson's book is sufficiently reliable, stable, and well documented, as a Harvard thesis deposited in the Harvard University Archive, to be given priority. Hence I feel that the priority of separability, sure-thing principle, independence, and all that, now lies with Samuelson rather than with Sono or Leontief. The priority may end up with authors preceding Samuelson, because the condition of cross-derivatives being independent of other commodities (well-known to be a preference condition: it amounts to the exchange rate between two commodities being independent of the other commodities and, hence, to separability of every pair of commodities from the other commodities) and its implication of additive representation was a kind of folk theorem. But with the evidence available now, for me Samuelson has the priority. I write this in my letter to him.
  5. Letter of Samuelson of 8 April 1996.
    Thanks me for my letter of 26 March 1996, and gives some after-thoughts.