A capacitor is consists of two conductors separated by a dielectric - that's the basic definition of a capacitor.
A trace on a PCB is a conductor; the material of the board is a dielectric. Any other trace or plane on the board that is in close proximity to a trace then becomes the second conductor, and the trace then has measurable capacitance.
All PCB traces have self-inductance, and mutual inductance with other traces or structures on the board. All PCB traces have some capacitance - the amount depends on the total area (length x width) of the trace, and the area and proximity of adjacent traces or planes.
There is nothing magic about three inches, or any other arbitrary length. The capacitance of a trace depends on what else is on your board, and how close it is to your trace.
A PCB trace is only one-half of the overall signal path. Every signal must have a return side - that forms a complete signal loop. The larger the loop is (the further away you make the return current flow), the larger the inductance of your trace will be. An ideal board has a return path directly below each signal trace. Normally this is done with a plane layer (either ground or power). As soon as you place that return path, you have increased the capacitance of the trace - BUT, at the same time you have drastically reduced the inductance of the trace by decreasing the signal loop size. Like everything in engineering, there are tradeoffs when designing a printed circuit board.
By controlling the width of the trace, and the distance between the trace and its return path, you can control the impedance of the trace for a given PCB dielectric material. The length of the trace only becomes important when you have signal timing issues that need to be controlled.