Potential explanations of the anomalous continuous absorption coefficient in water vapors (ACA) are examined theoretically. We show how the single line shape is determined by the time dependence of the dipole-dipole autocorrelation function(DDA) whose short time behavior dominates the frequency dependence in the far wing region. Based on general theoretical argument, it is shown that the DDA must be a Gaussian function for short times and an expotential function for long times. Empirical formula are obtained to fit the experimental data of CO2, whose absorbtion coefficient diminishes much faster than the Lorentz line shape. Since it is difficult to understand how CO2 or H2O can be different in the far wing region, we have to conclude that the ACA cannot be explained by modified local Lorentzian lines of H2O in the far wing region. We point out that both H2O and HCl which have permanent dipole moments have ACA. The possibility of the dipole-dipole interaction that may be important for ACA is examined.