QCD Regularization & Oriented Poles

Quantum chromodynamics — the theory of the strong force — is arguably the most precisely tested physical theory we have. Yet at its computational core sits a procedure that standard mathematics cannot fully justify: regularization, the controlled navigation of divergences at the boundary of the number system.

MMP's oriented void algebra offers a natural language for what is actually happening. Before ascending to full QCD, it is worth grounding the ideas in the simplest possible case.

The 0-dimensional toy model

Strip away all of spacetime. A 0-dimensional quantum field $\phi$ is a single variable — no position, no time, just a value. The "path integral" collapses to an ordinary integral, and the probability density is simply:

$$P(\phi) = N \, e^{-S(\phi)}$$

where $S(\phi)$ is the action (for a quartic theory, $S(\phi) = \frac{m^2}{2}\phi^2 + \frac{\lambda}{4!}\phi^4$) and $N = 1/Z$ is the normalization factor that makes the total probability integrate to one:

$$Z = \int d\phi \; e^{-S(\phi)}$$

This is the partition function — the total weight of all possible field values.

The MMP reading

$\phi$ in 0D has zero length: it occupies no spatial extent whatsoever. Yet it carries a full probability distribution — it has precision through $P(\phi)$. This is the $\{\}_+$ end of the MMP number line: zero length, non-zero precision.

The normalization $N = 1/Z$ is equally stark. $Z$ is the integral over all values of $\phi$ — the full "length" of the space of possibilities. $N$ is its reciprocal: as the space of possibilities grows ($Z \to \infty$), the normalization shrinks ($N \to 0$). Length and precision trade off exactly as MMP predicts.

The action $S(\phi)$ is the weight assigned to each value of $\phi$. It peaks where the field is most probable, falls off at the tails. The poles of the theory — where $e^{-S(\phi)}$ diverges or where $Z$ fails to converge — are exactly the points where the field tries to sit at $\phi \to \pm\infty$. These are the oriented poles $\{\}_-$ and $\{\}_+$ re-entering through the unbounded tails of the distribution.

In 0D there are no loop momenta, so there are no UV/IR divergences of the momentum-integral type. But the boundary structure is already visible: the partition function is finite only when $S(\phi)$ grows fast enough to tame the tails — only when the poles are oriented correctly and the integral can close.

The divergence problem

Loop integrals in QFT diverge at two boundaries:

• Ultraviolet (UV): $k \to \infty$ — the integral blows up at high momentum
• Infrared (IR): $k \to 0$ — the integral blows up at zero momentum

These are not symmetric accidents. The Heisenberg uncertainty relation

$\Delta x \cdot \Delta k \sim 1$

makes the trade-off explicit: push momentum to infinity ($k \to \infty$) and position localises to a point — zero length, infinite precision, $\{\}_+$. Push momentum to zero ($k \to 0$) and the state delocalises entirely — infinite length, zero precision, $\{\}_-$.

The UV and IR divergences are therefore the two oriented poles of the MMP number line expressing themselves physically. The full quaternary picture is present:

$$0 \circlearrowright \infty = \{\}_+ \qquad \infty \circlearrowright 0 = \{\}_-$$

Zero precision and zero length held in tension — the same quaternary balance that bequeaths the continuum, now showing up as the boundary conditions of quantum field theory. Standard arithmetic cannot evaluate at either pole; regularization is the controlled algebra of approaching without arriving.

The $i\epsilon$ prescription as orientation

The Feynman propagator writes the dangerous denominator as:

$$\frac{1}{k^2 - m^2 + i\epsilon}$$

The $+i\epsilon$ shifts the pole off the real axis, specifying which side to approach from when integrating over $k$. Remove it and the integral is undefined; restore it and the result is determinate.

MMP reads this directly: $i\epsilon$ is the directional information that the unoriented $0$ was missing. The propagator pole is a mystical number — $\{\}$ without orientation — and the $i\epsilon$ prescription orients it, choosing $\{\}_+$ or $\{\}_-$ before the composition is evaluated. Indeterminacy is the price of forgetting orientation; the $\epsilon$ is how physics quietly reinstates it.

Dimensional regularization as fractional demotion

A more powerful technique analytically continues spacetime from $d = 4$ to $d = 4 - 2\epsilon$, computes in the fractional dimension where integrals converge, then takes $\epsilon \to 0$. The divergences resurface as poles in $\epsilon$:

$$\frac{1}{\epsilon}, \quad \frac{1}{\epsilon^2}, \quad \ldots$$

These are then subtracted in the renormalization step.

In MMP's language: the fractional dimension step sidesteps the mystical boundary by never quite arriving at $d = 4$. The poles in $\epsilon$ are the oriented voids re-emerging as $\epsilon \to 0$. Renormalization — subtracting the pole — is an instance of the demotion operator $\circlearrowleft$: stripping one level of divergence to recover a finite residue.

$$\text{(divergent integral)} \;\circlearrowleft\; \{\}_\pm \;=\; \text{(finite result)}$$

The half-dimension and evanescent operators

In dimensional regularization, certain operator structures — called evanescent operators — vanish exactly in $d = 4$ but contribute in $d = 4 - 2\epsilon$. They are objects that only exist in the fractional-dimensional intermediate step, leaving a finite trace after $\epsilon \to 0$.

MMP would frame these as transiently inhabiting the void level: they are not physical at rank 1 ($d=4$) but they pass through the boundary and deposit something real before vanishing. The half-dimension is not a pathology — it is the oriented crossing point.

TODO

• Work out the specific MMP rank assignment for UV vs IR poles and their demotion residues
• Examine whether the renormalization group flow (running coupling) has a natural reading as iterated $\circlearrowright / \circlearrowleft$ compositions
• Investigate zeta function regularization ($\sum n = -1/12$) as another instance of analytic continuation through the mystical boundary