For what it's worth, Gerard Dumenil and my proposal to calculate the
value of money (or its inverse, the monetary expression of value) as the
ratio of realized value added to productive labor time was an attempt to
address precisely this question, since it does not make any assumption
about the forces determining relative prices.