Peter's last posting reminded me of something that may or may not be relevant. Sometime in the previous millennium (actually, around 3 decades ago) John Isbell made an observation that amounted to the statment that the equational theory of the unit ball functor of banach spaces (which has many more algebras than banach spaces) could be described by negation and an aleph_0-ary operation that takes {x_i} to \sum_{i=1}^\infty 2^{-i}x_i (and appropriate equations). Now a midpoint algebra with involution, as described by Peter, has all such finitary sums and if you also assume it complete, I think it is likely exactly a model of the banach space theory. Michael
Re Mike Barr's comments on Isbell's construction of unit balls, and Vaughan's remarks on Conway's construction, there is a paper by Denis Higgs (Proc. Koninklijke Nederlandse Akademie, Series A, Vol. 81, (4), 1978, pp.448-455) called: "A Universal Characterization of [0,\infty]", in which he gives such a characterization, based on a class of infinitary algebras in which the infinitary operation arises from the observation that every element of [0,\infty] can be wrtten as a sum, in general infinite, of fractions 1/{2^n} 's. Indeed, as Higgs' shows, this characterizes [0,\infty] as a free algebra of the appropriate kind on one generator. Cheers, Phil Scott
participants (2)
-
Michael Barr -
P. Scott