Intro to Renormalization Group

Introduces coupling flows, beta functions, fixed points, and critical scaling.

Conceptual context

Microscopically different systems can display the same behavior near a continuous transition. Renormalization explains this by changing resolution: short-scale details are removed, the description is rescaled, and its parameters are updated. The $g$ therefore follows a trajectory as $l$ changes. The $\beta_{g}$ describes that trajectory, while $g_estrella$ marks a description reproduced by scale change. The central issue is not merely the numerical interaction strength, but which information remains visible from farther away.

🟢 Essential level

Imagine repeatedly blurring an image. Fine details disappear at each step, while parameters are adjusted so that the visible physics remains correct. The $g$ expresses interaction strength at the current resolution, so it may change although the material system does not. The $l$ counts rescaling steps. The $\beta_{g}$ gives the direction and local speed of the flow. When the trajectory approaches an $g_estrella$, large-scale observations forget many microscopic details. The delta_g_0 measures the starting displacement from that destination, and the $y$ decides whether this displacement grows, shrinks, or requires finer analysis. With increasing scale directed toward long distances, negative $y$ erases the perturbation, positive $y$ makes it relevant, and a zero value is marginal.

This classification explains why microscopic differences become invisible while a few tuned controls still determine the observed phase.

Kenneth Wilson developed this renormalization-group formulation for critical phenomena in 1971, explaining for the first time why materials as different as a magnet and a fluid near their critical points share the same critical exponents. He received the 1982 Nobel Prize in Physics for that work.

🔵 Formal level

Basic formulas

Let $l$ be dimensionless and increase toward long scales. Near an $g_estrella$, the $g$ is written as the fixed value plus a small deviation. Its linear evolution is

Linearized coupling flow

where delta_g_0 gives the deviation at the reference scale and $y$ is the linearized eigenvalue. Negative $y$ produces exponential decay, positive $y$ produces growth, and zero leaves the linear analysis undecided.

Real student questions

Is the group an ordinary symmetry group? Modern coarse graining may form only a semigroup because discarded information is not recoverable.

Must a fixed point have zero coupling? No. The Gaussian point is often zero, but interacting fixed points exist.

Can $\beta_{g}$ be small away from a fixed point? Yes. Slow parametrization or higher-order structure can make it small; locate its zero and test stability.

Is scale a time variable? No. $l$ labels resolution. The dynamical analogy is mathematical.

Final synthesis

Renormalization turns resolution change into parameter flow. $\beta_{g}$ directs $g$, $g_estrella$ expresses scale invariance, and $y$ classifies surviving perturbations. Every interpretation must state the scale convention and verify the local validity of linearization.