Re: "manifold" in mathematics

Posted by Steve Smith on
URL: http://friam.383.s1.nabble.com/manifold-in-mathematics-tp3385914p3393427.html

[hidden email] wrote:
Let me add another inquiry to this - how do we reconcile this notion of manifold with the idea of self-similarity? If Epping Forest is a manifold, but the leaves and twigs are not, yet the leaves and twigs have some self-similarity, is Holt truly thinking in terms of the mathematical definition of manifold, as Roger gave us, or is the metaphor missing something (or am I)?
Clay, et al. -

I would submit that the term (and especially concept) of manifold significantly predates the mathematical use.   This concept and usage would seem to have provided the motivation for the mathematical term itself.   It also might provide a little more leeway in expanding the application to reference or accommodate self-similarity in some way, as you suggest.

From: English Etymology Dictionary

    manifold O.E. monigfald (Anglian), manigfeald (W.Saxon), "varied in appearance," from manig "many" + -feald "fold." A common Gmc. compound (cf. O.Fris. manichfald, M.Du. menichvout, Swed. m?ngfalt, Goth. managfal?¢®s), perhaps a loan-translation of L. multiplex (see multiply).

And from: Webster's Revised Unabridged Dictionary (1913)

    Manifold \Man"i*fold\, a. [AS. manigfeald. See Many, and Fold.] 1. Various in kind or quality; many in number; numerous; multiplied; complicated. O Lord, how manifold are thy works! --Ps. civ. 24. I know your manifold transgressions. --Amos v. 12. 2. Exhibited at divers times or in various ways; -- used to qualify nouns in the singular number. ``The manifold wisdom of God.'' --Eph. iii. 10. ``The manifold grace of God.'' --1 Pet. iv. 10.


My working definition (sharpened in the process of this discussion, thank you Nick!) is to "be many and varied in nature" with some alternate but related working definitions. "a mechanical multiplexer, as in a gas manifold such as used for intake fuel mixture and exhaust gas in an internal combustion engine."  This usage seems to reference the complex topology requiring something of a patching together of various tube-shaped surfaces". and "multi-fold, in the sense perhaps of complex origami pieces or even the simplicity of carbon-paper copying... or multifold writing".

As for your suggestion...   I submit that for my purposes, geometric artifacts exhibiting self-similarity (often fractals) do in fact suggest a valuable variation on the use of "manifold".    One might say that such artifacts have sub-components which are "many and varied in nature", though the definition of "varied" is challenged a bit by the self-similarity.   They are clearly varied (being similar, not identical) yet, are not-so-varied (being similar, not different).  Fractals (and other self-similar geometries) embedded as surfaces in 3 dimensional spaces do strike me intuitively as being "barely" or "almost" a manifold.   They are "barely" in the sense (again) of their subcomponents (regions) being many and vairied but "almost" in the sense of not being as diverse as I would normally want.   

A *real-world* self-similar system such as your referenced _forest of trees of branches of twigs with leaves having at least a few levels of self-similar structure_ would seem to be very apt and perhaps where some of the earliest referenced usage of the *concept*  ("O Lord, how manifold are thy works!") arose.

<nostalgic anecdote>
My first encounter with the term was on the 390 CI engine of my 1964 Ford Thunderbird in High School.   I confronted my learned (aka geek-nerd) friend whose father was an engineer and who bent and welded up his own headers for the ford 289 CI in his 1963 Ford Fairlane coupe with: "what is the difference between a manifold and a header?".   His response was somewhat like Doug's or Owen's to many of Nick's questions here, but I persisted until his father overheard the conversation and weighed in himself.   As an engineer he was more interested in the practical properties rather than what I was seeking of the "essential" properties and proceeded to explain to me all about pulse-tuning of the exhaust system, how impedence worked in complex, dynamic systems of compressible fluids (exhaust gas in particular) down to how the cooling of the exhaust gasses in the exhaust system effected the pressure/volume and how that should be accounted for in the tuning even.   Once we got past the pragmatics, I was able to engage him in my *real* question which was why a welded-up header was not also a manifold (even then I knew how to read a dictionary and had at least been introduced to non-euclidean geometry and topological spaces).  What we settled on (so that my friend's mother didn't kill us all for holding up dinner with this long-winded conversation) was quite obvious, as the headers were an array of exhaust pipes connected for convenience by a single flange for bolting up in place of the manifold.   We agreed that perhaps the *other end* of the header-system where they all joined to a single exhaust pipe (well 4-1 for each of 8 cylinders and 2 exhaust pipes) was a bit of a "manifold" in the original sense but that since the joints were roughly non-differentiable, not really.   They were a "multiplex" of 2D surfaces joined in a complex way, but not (quite) a manifold in the more strict sense.  
</nostalgic anecdote>

I shall now return to my manifold deadlines and projects (all joined together in a  singular and continuously differentiable high-dimensional artifact embedded in a higher dimensional space)

- Steve

PS.  If you read this far Gattiker, drop me an e-mail!

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org