A couple of things. First, I neglected to mention that "Communes via Yoneda, from an Elementary Perspective," Fundamenta Informaticae 123 (2010) 1–16, DOI 10.3233/FI-2010-315 is about to appear and won't be findable by Google just yet. Those interested in seeing it sooner can find it on my site at http://boole.stanford.edu/pub/CommunesFundInf2010.pdf Second, as I said I wasn't passing judgment on the wisdom of avoiding the term "schzophrenic" but merely pointing out the associated cost, which needs to be balanced against the harm of any given word. So I followed Tom's pointer http://ncatlab.org/nlab/show/dualizing+object linking to a discussion of alternatives, which seemed inconclusive. Sam (Staton?) made the point however that even if schizophrenia is not the appropriate word, schizo is the appropriate prefix, having derived from the Greek "split." So it is the medical condition that is inappropriately named, namely as "split madness," with phrenitis and frenzy having a common origin. With that in mind it occurred to me that "schismatic" might be a suitable alternative, as providing better continuity with the older terminology by coming from the same root schizo, but more honestly so than schizophrenia since in this case there really is a multiple personality, and moreover there's nothing insane about it. (And it's a syllable shorter to boot.) Third, while it is true that the schismatic object (to give the term a trial run) is usually observed manifesting its split personality in different categories, this is not the case in *-autonomous categories where I and _|_ are the Jekyll and Hyde of the same category. (I apologize to readers of this list with either of those surnames.) In all the examples I'm aware of, the two categories in which the schismatic object occurs (once in each) admit a common completion to a *-autonomous category which embeds one object as I and the other as _|_. Considering them to be the "same" object found in two categories misses the contravariance between them, which is brought out more clearly by this joint completion, where they are clearly not the same object but a pair of dual objects. My paper accounts for C.I. Lewis's qualia by viewing them as morphisms running from I to _|_. If I and _|_ are rigid (|C(x,x)|=1) as for Chu(Set,K)), the presence of a morphism from _|_ to I is inconsistent in the sense that it collapses Hom(I,_|_) to a singleton, since I and _|_ respectively generate and cogenerate. So in order to have more than one quale (in the Chu setting) there cannot be any morphism from _|_ to I. (That was mainly in the nature of background on the neighborhood of I and _|_.) Vaughan [For admin and other information see: http://www.mta.ca/~cat-dist/ ]