The lack of a reliable theory of glass physics has led to the pursuit of correlations between various glass or viscous liquid parameters, one of which is the slope m of the plot of log(viscosity) against Tg/T, extrapolated at the glass-transition temperature, Tg, also termed 'fragility'. Novikov and Sokolov conclude that the value of m for a liquid varies linearly with the ratio of the instantaneous bulk and shear moduli, K infinity/G infinity, of its glass according to the relation m=29(K infinity/G infinity -0.41). Because of the obvious importance of the elastic properties of a glass, we have investigated the basis for this relation and find that its premise is flawed because of the unjustifiable preference for an empirical variation of m with elastic properties, and because of the selected use of glasses. When more glasses are considered in the same way, m does not seem to be linearly related to K infinity/G infinity.