You probably learned in middle school that a triangle's angles always add up to 180 degrees. On top of that, two right angles would eat up the whole budget. Which means third angle: zero degrees. Not a triangle. Case closed.
Except the case isn't closed. Not even close That's the part that actually makes a difference..
What Is a Triangle, Really?
Here's the thing — "triangle" isn't a single, universal object. It's a definition that changes depending on what surface you're drawing on.
In Euclidean geometry — the flat-plane geometry Euclid codified around 300 BCE — a triangle is three straight line segments connecting three non-collinear points. The angles sum to exactly 180°. That said, always. Two right angles? Because of that, impossible. The third side would have to be a line of zero length, which means the two vertices collapse into one. You get a line segment, not a triangle.
Not obvious, but once you see it — you'll see it everywhere.
But Euclid's fifth postulate — the parallel postulate — only holds on a flat plane. Real worlds aren't flat Easy to understand, harder to ignore..
Triangles on a Sphere
Grab a globe. Put your finger on the North Pole. Trace a line straight down the prime meridian to the equator. In practice, turn 90° left. Walk along the equator a quarter of the way around the planet. Turn 90° left again. Walk straight back to the North Pole Small thing, real impact..
This is where a lot of people lose the thread Worth keeping that in mind..
You just drew a triangle with three right angles. 270° total.
That's not a trick. That said, that's spherical geometry — the geometry of the surface of a sphere. Which means the "lines" are great circles (the largest possible circles on the sphere, like the equator or lines of longitude). They're the straightest paths available on a curved surface. Geodesics, if you want the technical term.
On a sphere, the angle sum of a triangle is always* greater than 180°. It can go all the way up to 540° for a triangle that covers almost the entire sphere. Two right angles? And easy. Three right angles? Standard. The "excess" over 180° is directly proportional to the triangle's area — a beautiful result called Girard's theorem It's one of those things that adds up..
Triangles on a Saddle (Hyperbolic Geometry)
Flip the curvature the other way. In practice, a hyperbolic plane — think of a saddle shape or a Pringles chip, extended infinitely — obeys a different rule. That said, triangle angles sum to less* than 180°. Consider this: you can have triangles with two angles arbitrarily close to 90° and the third angle tiny, but never two exact* right angles with a positive third angle. The defect (180° minus the sum) is proportional to area here too It's one of those things that adds up..
So the answer to "can a triangle have two right angles" is: on a flat plane, no. On a sphere, yes — and it can have three.
Why This Matters (More Than You Think)
This isn't just a party trick for math nerds. It's how GPS works. On the flip side, it's how airlines plan routes. It's why the shortest flight from New York to Tokyo arcs over Alaska instead of cutting straight across a flat map.
Navigation and Great Circles
Every long-distance flight path you've ever seen on a seatback screen — that curved arc — is a segment of a great circle. But it's a "straight line" on a sphere. Pilots and shipping routes follow geodesics because they're the shortest distance between two points on Earth's surface.
If you treated Earth as flat (Euclidean), you'd calculate the wrong distances, the wrong headings, and you'd burn extra fuel. The difference matters. A great-circle route from London to San Francisco saves hundreds of miles compared to a rhumb line (constant compass bearing) — which looks straight on a Mercator map but curves on the globe Most people skip this — try not to..
Map Projections Are All Lies
Every flat map distorts something. Mercator preserves angles but explodes area near the poles. Practically speaking, equal-area projections preserve relative size but mangle shapes. The Peters projection, the Robinson, the Winkel tripel — they're all compromises.
Understanding spherical triangles explains why no flat map can be perfect. Which means you're trying to flatten a geometry where triangles sum to >180° onto a geometry where they sum to exactly 180°. Something has to break. Angles, areas, distances, shapes — you pick your poison Surprisingly effective..
Not the most exciting part, but easily the most useful.
General Relativity Takes It Further
Einstein didn't stop at spheres. That's why in general relativity, spacetime itself is curved by mass and energy. Light follows geodesics in this curved 4D manifold. Now, the angles of a triangle formed by three laser beams in deep space? 180°. Near a black hole? More. The curvature of space is literally measured by how much triangle angle sums deviate from the Euclidean expectation Small thing, real impact. Worth knowing..
Gravitational lensing — the bending of starlight around the sun that made Einstein famous in 1919 — is exactly this phenomenon. The "triangle" formed by the star, the sun, and the observer has an angle sum that doesn't match flat-space predictions.
How Spherical Triangles Actually Work
Let's get concrete. You don't need a globe to see this — you need the right mental model.
The Building Blocks: Great Circles
A great circle is any circle on a sphere whose center coincides with the sphere's center. Every line of longitude. The equator. Any circle that cuts the sphere into two equal hemispheres The details matter here..
Small circles (like lines of latitude except the equator) are not geodesics. Consider this: if you walk along the 45th parallel, you're constantly turning slightly — you're not following a "straight line" on the sphere. Only great circles are locally distance-minimizing paths.
Defining the Angles
On a flat plane, the angle between two lines is unambiguous. Day to day, on a sphere, two great circles intersect at two antipodal points. At each intersection, they form four angles (two pairs of vertical angles). The interior angle of a spherical triangle is the angle between the tangent vectors* of the two great circles at the vertex, measured inside the triangle It's one of those things that adds up. Nothing fancy..
We're talking about why the North Pole triangle works. The prime meridian and the 90°E meridian meet at the pole at a 90° angle. That said, the equator meets each meridian at a 90° angle. Three 90° angles. 270° total.
Area and Angle Excess
Girard's theorem (sometimes called the spherical excess formula) is elegant:
Area = R² × E
Where R is the sphere's radius and E is the spherical excess in radians: E = (A + B + C) - π That's the whole idea..
For the three-right-angle triangle on a unit sphere (R=1): E = 3π/2 - π = π/2. That said, area = π/2 — exactly one-eighth of the sphere's surface area (4π). Makes sense: the triangle covers one octant of the globe It's one of those things that adds up..
This relationship — angle excess proportional to area — is a hallmark of positive curvature. In hyperbolic geometry (negative curvature), it's angle defect* proportional to area. In Euclidean geometry (zero curvature), excess and defect are both zero — area is independent
This changes depending on context. Keep that in mind.
of the angle sum. Practically speaking, on a flat plane, a triangle the size of a coin and one the size of a continent will both have interior angles adding up to exactly 180°. The angles tell you nothing about the area, because the underlying space doesn't curve.
The Hyperb
The Hyperbolic Triangle
In spaces of negative curvature, the familiar 180° rule flips on its head. Think about it: a hyperbolic triangle—formed by three geodesics on a surface that curves away from itself—always has an interior‑angle sum strictly less than π radians (180°). The shortfall, called the angle defect, is not a quirk; it is directly proportional to the triangle’s area, just as the spherical excess is for positively curved surfaces.
The relationship is elegantly simple:
[ \text{Area} = R^{2},\bigl(\pi - (A+B+C)\bigr), ]
where (R) is the (absolute) curvature radius of the hyperbolic plane (for the standard model with curvature (-1), (R=1)). The larger the defect, the larger the triangle’s area. In the limit, a triangle can be arbitrarily huge while its angles remain tiny, a hallmark of the “expanding” nature of hyperbolic space.
Visualising Hyperbolic Geometry
Because we live in a Euclidean world, we rely on mathematical models to “flatten” hyperbolic space for intuition:
-
Poincaré disk model – the entire hyperbolic plane is represented by the interior of a unit circle. Geodesics appear as arcs of circles that intersect the boundary orthogonally, or as diameters. Distances swell dramatically near the edge, so a tiny Euclidean segment near the boundary corresponds to a long hyperbolic path Worth keeping that in mind..
-
Upper‑half‑plane model – the set ({(x,y)\mid y>0}) with the metric (\frac{dx^{2}+dy^{2}}{y^{2}}). Geodesics are either vertical lines or semicircles centred on the real axis. This model is especially convenient for studying modular groups and tessellations.
Both models preserve angles (they are conformal), which makes them ideal for discussing triangle angles, yet they stretch lengths in a way that reproduces the constant negative curvature Not complicated — just consistent..
A Classic Example: The Ideal Triangle
Consider a ideal triangle, whose three vertices lie on the boundary of the model (at infinity in the Poincaré disk). Remarkably, every ideal triangle in hyperbolic space has the same area, regardless of how “large” it appears in the Euclidean picture. Its area is therefore (\pi) (in units where (R=1)). Day to day, each interior angle is zero, so the defect is (\pi). This universality is a striking contrast to spherical triangles, where area grows with excess, and Euclidean triangles, where area is independent of the angle sum.
Why the Defect Matters
Just as spherical excess tells us how much a surface bends outward, hyperbolic defect tells us how it bends inward. That said, in both cases, the deviation from the Euclidean 180° sum is a direct geometric measurement of curvature. On a sphere, excess is positive; on a saddle‑shaped hyperbolic plane, defect is positive (or excess negative) Not complicated — just consistent..
[ \text{Curvature } K = \frac{E}{R^{2}} \quad\text{(spherical)}\qquad K = -\frac{D}{R^{2}} \quad\text{(hyperbolic)}, ]
where (E) is excess and (D) is defect Easy to understand, harder to ignore..
Connecting to the Cosmos
Gravitational lensing, as we saw, probes the curvature of spacetime along the light’s path. In regions of positive curvature
Gravitational lensing provides a direct experimental window onto the curvature described above. On the flip side, light rays travel along null geodesics that are bent by massive objects exactly as predicted when the underlying space possesses a non‑zero Gaussian curvature. In an expanding universe, the sign of that curvature determines whether the bend is amplified or suppressed. To give you an idea, in a positively curved (closed) cosmos, the spatial sections have finite volume and the effective deficit angle measured around a closed orbit would be less than (2\pi); conversely, in a negatively curved (open) universe the same observation yields a surplus angle greater than (2\pi). Modern surveys of galaxy clusters and weak‑lensing shear maps consistently infer a small but definitive value of the curvature parameter (\Omega_{\text{k}}<0), confirming a flat‑to‑open geometry at the scale of cosmological horizons The details matter here..
The mathematical link is transparent once we translate the curvature into the defect language introduced earlier. The defect (D) of a hyperbolic triangle equals (-K R^{2}), where (K) is the (negative) Gaussian curvature. Think about it: thus a region of space with curvature (|K|) produces a measurable excess defect whenever one draws sufficiently large triangles, and the size of that excess scales linearly with the square of the radius used. By embedding a patch of the actual Universe—say the local supercluster surrounding the Milky Way—in the hyperbolic model, observers would predict a systematic over‑compensation of angular deficits compared with the Euclidean expectation, a prediction that current data already support within uncertainties.
Beyond cosmology, the same principle underlies many other physical contexts. So naturally, in condensed matter physics, defects such as dislocations behave analogously to hyperbolic triangles: the local curvature generated by a line defect forces nearby lattice sites into a net outward flow that mimics the “saddle” shape of hyperbolic space. Numerical simulations that incorporate curvature corrections often reproduce this effect through a term proportional to the defect, providing a bridge between abstract geometry and material science Less friction, more output..
Returning to the original discussion, the capacity of hyperbolic geometry to accommodate infinitely large figures while preserving infinitesimal angle measures offers a powerful mental image for thinking about spaces that are far more expansive than our everyday experience. The universal area of an ideal triangle ((\pi)) illustrates that the intrinsic geometry does not depend on how “big” an object looks in a Euclidean projection; instead, it is governed solely by the global curvature signature encoded in the defect. This invariance is what makes hyperbolic space a fertile arena for exploring the interplay between topology, curvature, and observable phenomena Simple, but easy to overlook..
In a nutshell, the defect‑based characterization of curvature bridges pure mathematics with concrete astrophysical observations. By recognizing that the curvature of the cosmos leaves an imprint on the way light bends and on the total area of maximally symmetric triangles, we gain a unified framework that links the elegant theory of hyperbolic geometry to the very fabric of the universe itself Still holds up..
Worth pausing on this one.
