We show how to use 21 cm emission and absorption studies to estimate the heat inputs to the neutral gas in low pressure environments, such as in outer disks of spiral galaxies, in galactic halos or in intergalactic space. For a range of model parameters we calculate the gas kinetic and spin temperatures (T(K) and T(S)) and the relation between T(S) and the heat input to the gas. We outline the conditions for a ''two-phase medium'' to exist. We find that although T(S) can be much smaller than T(K), T(S) is always much greater than 3 K for column densities greater than 5 x 10(18) cm-2. This excludes the possibility that relevant H I masses at the periphery of galaxies are invisible at 21 cm in emission. Therefore sharp H I edges, observed in outer disks near column densities N(l) approximately 2 x 10(19) cm-2, cannot be ''fictitious'' edges due to a sudden decrease of the 21 cm brightness. The outermost interstellar gas in a disk galaxy is more directly affected by external processes, and in this paper we estimate the intensity of the extragalactic background at energies close to 0.1 keV by comparing our theoretical results with H I emission/absorption studies. We take into account the possibility that some energy produced in the inner regions affects the energy balance in outer regions. We find that in the absence of any other local heat source QSO-dominated background models are still compatible with the spin temperature limits derived for the two best documented H I emission/absorption studies in outer regions. However, if future observations should establish that the spin temperature is as high as 1000 K, then relevant energy inputs from local sources become necessary.