it might help a bit if you knew the archaic name of a diff amp, "emitter-coupled amplifier". the emitters are at the same potential, so what happens on one side has an effect on the other side. the resistor feeding the emitters is usually not a resistor, but a constant current source, and the output from the collectors is usually in the form of a current and not a voltage. with a constant current, the two transistors have equal currents through them "at rest". if you increase the current through Q1 by driving it's base with more current, more current flows through Q1. that additional current raises the voltage on Q1's emitter, which is directly connected to Q2's emitter. when Q2's emitter also goes higher, which reduces the bias on Q2's base by the same amount, which reduces Q2's current. the total current remains the same. think of it as a seesaw with a fulcrum in the middle.