geometric group theory
A group is an algebraic gadget: a set of symmetries or operations you can combine, like the moves of a Rubik's cube or the integers under addition. For a long time groups were studied purely as algebra. Geometric group theory is the modern, beautiful turnaround: it studies a group by turning it INTO a geometric shape and reading off the group's properties from the shape's geometry. The slogan is that you can understand an abstract group by looking at the 'space' it secretly is.
Here is how a group becomes a space. Choose a set of generators (basic moves from which all others are built). Draw a dot for every element of the group, and connect two dots by an edge whenever one is obtained from the other by a single generator. The resulting network is the Cayley graph, and it turns the group into a connected geometric object with distances: the distance between two elements is the fewest generators needed to get from one to the other (the 'word metric'). Now the group HAS a shape, and you can ask geometric questions about it. Does it look like a flat grid, like the integers in two directions? Does it spread out exponentially like a tree, the signature of 'negative curvature'? How fast does the number of elements within distance n grow — its growth rate? Mikhail Gromov's deep insight was that large-scale geometric features of this shape (coarse properties that ignore local detail, captured by 'quasi-isometry') reflect genuine algebraic facts about the group, and conversely.
This viewpoint reshaped group theory and links it to topology, hyperbolic geometry, and computer science. Gromov's theorem that groups of polynomial growth are 'virtually nilpotent', and the rich theory of hyperbolic groups, are landmark fruits; the ideas reach into the study of 3-manifolds and even efficient algorithms on groups. It is firmly a FRONTIER subject, so honesty matters. The Cayley graph's fine shape depends on which generators you pick, so geometric group theory deliberately studies only the COARSE, large-scale geometry that survives that choice (quasi-isometry invariants) — the small-scale picture is not intrinsic to the group. And while the field has spectacular successes, many basic questions remain open; it is an active, unfinished area, not a closed catalog of theorems.
Take the simplest infinite group, the integers under addition, with the single generator +1. Its Cayley graph is just an endless straight line of dots ..., -2, -1, 0, 1, 2, ... each joined to its neighbors. The word distance between 0 and 5 is 5 — you need five steps of +1. So this group, viewed geometrically, IS a line: 1-dimensional, flat, growing linearly. Take instead the integer grid in two directions and its Cayley graph is a flat square lattice, growing like n^2 — and that quadratic growth is a geometric fingerprint of the algebra.
A group becomes a shape via its Cayley graph, and the shape's growth rate echoes the group's algebra.
Only the LARGE-scale geometry of the Cayley graph is intrinsic to the group; the fine detail depends on which generators you chose. Geometric group theory therefore works with coarse, quasi-isometry-invariant features, not the local picture.