Dear Lawrence,
This is response to your post of Sept. 29, 03:25 GTM.
Let me begin with a piece of reverse advertisment concering my book. If you are now working on something else and intend to get to quantions later, I would suggest you wait and then contact me before looking for the book. I might have an update by then. The book is not 'wrong', it's just 'old fashioned'. For an animal as young as quantions, two years is not not a short time.
In my essay, I emphasized the importance of a phase I referred to as "polishing", or "finalization" (I don't remember which word I used). Briefly, a new theory is usually formulated in the mathematical formalism the author has at hand, but before going ahead with it, it makes sense to develop a formalism that does most of the work for us. After a couple of passes, I now have a 'sufficiently optimal' one for quantions. For example, the derivation of the quantionic version of Dirac's equation using matrix algebra is spread out over 40 pages in the book, but comes out of a few lines of linear algebra in the new formalism. This is not just about saving paper: It makes the deductive distance between the principle at the source of the quantionic Dirac equation (I call it "Zovko's intepretation") and the equation herself so short that it comfortably fits into our brain's working memory. We then have a much better understanding of what's going on and the correct intuitions for going ahead with new research. An since I am always afraid of having made an algebraic error, or of having overlooked something, it's quite a relief to see, as in the above example, that the new elegant derivation and the original messy one agree in their results.
********************
Yes, twistors and quantions are related --- the way two branches of a tree are related: They come out of the same trunk, but then have their own lives. This is worked out in detail in my first book ("The algebra of quantions" 2005). Here is a synopsis:
The geometric track:
Slightly rearranging history for the sake of conceptual coherence, we may say that Roger Penrose started with the conformal compactification of Minkowski space, whose invariance group is SO(2,4). Since the Lie algebras so(2,4) and su(2,2) are isomorphic, Penrose started investigating the pseudo-unitary space of quadruples of complex numbers {u,v,w.z} whose norm is defined as uu*+vv*-ww*-zz*. These are the original twistors. They are to the conformal group what Pauli spinors are to the rotation group.
The algebraic track:
The composition principle applied to the abstract structure extracted from classical mechanics yields, as a unique abstract solution, an algebra Petersen and myself called "quantal algebra". See:
"Inherently Relativistic Quantum Theory" Part I. The Algebra of observables. Fizika B (Zagreb) 10 113-138 (2001). On line at http://fizika.hfd.hr)
Given an abstract algebra, the next step is classification, which consists in finding all its concrete realizations. This is done in
"Inherently Relativistic Quantum Theory" Part II. Classification of solutions. Fizika B (Zagreb) 10 139-160 (2001). On line at http://fizika.hfd.hr).
The Lie algebras of the solutions are su(n), so(6), and so(2,4). There is nothing else (unless I made a mistake -- but probably not, because the proof, which is a modification of Cartan's proof, is not very complicated). The solution so(6) seems spurious to me, but it would be interesting if someone else saw in it something I missed. The solution so(2,4) brings us into the world of relativity. This is encouraging because the starting point was the abstract structure of observables. Not only that relativity was not postulated, geometry was not either. Yet, there is no mystery. An abtract simple Lie algebra makes no reference to geometry, yet Cartan's classification brings all spaces with a Pythagorean (or pseudo-pythagorean) metric to the light of day.
Given su(2,4), one can continue along the twistorial branch. But there is also another branch, which takes us from the conformal group SO(2,4) to the Lorentz group SO(1,3). This can trivially be done geometrically (by freezing two dimensions), but the quantal algebra offers a non-trivial option. It is a non-standard complexification possible only because the metric (+,+,-,-,-,-) admits an imaginary unit sqrt(-I) which is different from sqrt(-1). The latter leads to complex numbers, the former to quantions.
Coming back to your question concerning the quantion/twistor relationship, I just described their common algebraic source. They are brothers. Will they meet again in the future? I don't have the silghtest idea. If they do, I will be very pleased. It seems to me that they parted company like the brothers Brutus and Cassius in Shakespeare's Julius Caesar with the words:
And whether we shall meet again I know not.
Therefore our everlasting farewell take:
For ever, and forever, farewell, Cassius.
For if we do meet again, why, we shall smile;
If not, why then, this parting was well made.
********************
Quoting you:
" I think the composability requirement maybe satisfied for the exceptional algebra and its anti-Hermitian pair under G_2 holonomy."
If my proof is correct, no. But even at the cost of being embarrassed for having overlooked something, I would prefer it if you were right. It could be a source of some interesting physics. Beyond this, I have nothing useful to say.
Best regards, and thank you for bringing up this discussion.
Emile.