You shot at ISO 400 but needed ISO 800, so you told the lab to push it a stop. What you actually did wasn't make the film faster. Film's sensitivity to light is fixed at the moment of exposure — nothing that happens afterward in a tank of developer changes how many photons hit the emulsion. What extended development changes is how much of the silver that photons did hit turns visible, and how loudly.
The characteristic curve is where this plays out. It plots density against log exposure, and normal development draws a moderate slope through that relationship. Push processing extends the development time (sometimes the temperature too), and the straight-line portion of the curve steepens in response — a small difference in exposure now produces a larger difference in final density than it would have under standard development. That's the entire mechanism: you're forcing midtones to develop the way shadows normally would, buying back some of the density an underexposed frame is missing.
The Grain Doesn't Get a Vote
Extended development doesn't selectively work on the parts of the image you want more contrast in. It acts on every exposed silver halide crystal in the frame, and grain is nothing but the random spatial distribution of those crystals. Push the development and you push all of it — the crystals carrying real scene information and the crystals sitting in the sparse, barely-exposed shadow regions where the signal was thinnest to begin with.
That's exactly where grain becomes a problem rather than a texture. In a properly exposed midtone, there's enough silver developed that the random variation between individual crystals averages out into something that reads as smooth density with a fine overlay of texture. In an underexposed shadow pushed hard to compensate, there's very little silver to begin with, so the same random variation that used to average out now dominates what little density exists. The result isn't a fuller, cleaner shadow. It's a noisier, coarser one — the push didn't manufacture detail that wasn't captured, it just made the absence of detail visible as texture instead of as flat black.
Pushed Grain Isn't Native High-ISO Grain
It's tempting to treat a pushed stop as interchangeable with a native higher-speed film, but the two arrive at coarser grain through different routes. Native high-ISO stock uses physically larger silver halide crystals from the start — more surface area per crystal, which is what lets it reach a latent image threshold with fewer photons. Why ISO 400 is coarser than ISO 100 covers that mechanism directly: bigger targets, bigger resulting grain, by design.
Pushing doesn't touch crystal size at all. The crystals in a pushed roll are whatever crystals that stock always had — you're not making them bigger, you're developing them harder. What you get is the original crystal population under amplified contrast: a grain structure that reads as harsher and more clumped than that same stock's grain under normal development, but structurally distinct from what a stock with genuinely larger native crystals produces. Photographers who expect pushed ISO 400 to simply look like native ISO 1600 are often surprised by how different the two textures actually are under a loupe.
What Pushing Cannot Do
None of this recovers information the lens and emulsion never captured. If a shadow area received essentially no exposure, extending development doesn't discover detail hiding in it — there's no latent image there to develop, amplified or otherwise. Push processing is purely a density and contrast operation on silver that was already exposed. It cannot resolve an edge the lens didn't focus, and it cannot deposit silver where no photon landed.
That's the ceiling on how far pushing can go. Each additional stop demands more extended development, which steepens the curve further, which coarsens the grain further and compresses the highlights further — until the trade stops paying for itself and the image just looks blown out and gritty rather than merely fast. Two or three stops is a common practical limit for exactly this reason: past that point you're not buying usable speed anymore, you're buying an entirely different, harsher-looking negative.
The chemistry doesn't know what look you were hoping for. It develops whatever silver got exposed, at whatever rate the extended time and temperature drive it to. Everything past that — whether the resulting grain reads as gritty character or as a ruined shadow — is your call to make before you ever load the roll.