Torsten,
I finally got a little bit of time to write more on what I had mused about a couple of weeks ago. This all seems to center in a way around a type of cobordism with respect to these replacements of handles or Casson handles. The replacement of a circle with a knot suggests a type of theory that involves Hopf links. The trefoil for instance is by the Jones polynomial such that a left - right trefoil equals a Hopf link.
The manifold constructed from the knot K is
M_k = ((M^3\D^2xS^1)xS^1)∪_T^3 ((S^3\(D^2xK))xS^1).
On the left the R^1 in M^4 = M^3xR is replaced by S^1, and we can think of the S^1 as a periodic cycle with a real number line as a covering. Think of a wheel rolling on the real number line, or a spiral covering of a circle. In this setting the crux of the matter involves replacing a circle S^1 with a knot K. Physically this avoids topologies with circular time or closed timelike loops such as the Godel universe. The S^1 to the right of each expression is the embedding "time cycle" and the three manifolds of interest are (M^3\D^2xS^1) and S^3\(D^2xK). In a thin sandwich, a narrow section of spacetime separated by two spatial surfaces, we may think of the bottom spatial surface or bread slice as (M^3\D^2xS^1) and the second one as S^3\(D^2xK). We might further be so bold as to say the bottom surface is a left handed trefoil and there is a superposition of two surfaces, one with a right handed trefoil and the other with two S^1s in a link. There is then a type of cobordism between the bottom slice of bread and the top, which in this case might be a map from (M^3\D^2xS^1) ∪_T^3 S^3\(D^2xLT), for LT = left refoil to (M^3\D^2xS^1★S^1)∪_T^3 S^3\(D^2xRT). There the star means linking.
This is a theory of topology change in spacetime, or of some underlying topological change in topology which still maintains an "overall smooth" structure. This is then a type of topological quantum field theory (TQFT). A TQFT just means a theory that is a quantum field theory up to homotopy. This is a way of looking at fields (eg the knots as Wilson loops of fields) according to the underlying space they exist on. This approach amounts to cutting up the space into pieces, examining the fields there and then looking at the entire ensemble (pieces up back). This then has an underlying locality to it this way. However, the connection between knot polynomials and quantum groups indicates there is also something nonlocal as well.
This conjecture means that TQFT assigns data to all possible geometric element to a space, from a 0-dim point to the full manifold in an n-dim cobordism. For a space of n-dimensions there is a functor F
F:bord_n^f --- > A
For A an algebra. The algebra is the generator of the group G = quantum group. Physically the algebra corresponds to the connection coefficients A which form the Wilson loops ∮A•dx = ∫∫∇•Ada (to express this according to basic physics). This is a sort of Grothendieck topos or category system, which relates a knot group with a cobordism. I conjecture that a complete understanding of this system is a TQFT.
I will write in greater detail later on this, for I have sketched out some of this. Physically (or philosophically if you will) the description of spacetime this way is I think equivalent to a description of TQFT in general. In fact one result of the AdS/CFT correspondence is that a 4-spacetime as the boundary of an AdS_5 is equivalent to 10-dim supergravity. The exotic structure of 4-dim manifolds may then be a manifestation of 10-dim supergravity.
I copied this on my essay blog site, so if you respond to this there I get an email alert.
Cheers LC