Because of the impact of the El Nino-Southern Oscillation (ENSO) phenomenon on regional and global weather and climate, its potential long-term predictability is an important area of study. One plausible avenue of approach that has been attempted is to study the low-frequency modulation of the ENSO phenomenon. We consider here specifically the modulation of ENSO on decadal timescales by the Pacific Decadal Oscillation (PDO). Since each phase of the PDO has decadal persistence, it was hoped that its influence on the ENSO phenomenon might have decadal predictability. However, the nature of such an influence is not yet understood. There is no doubt that PDO can affect the tropical sea-surface temperature (SST), mostly in the equatorial Central Pacific, but, is this influence through the nonlinear mechanism of amplitude and frequency modulation of ENSO? Or is it through a linear superposition of the two climate modes on tropical SST? By showing that it is largely the latter, we suggest that the problem of decadal predictability should be recast into predicting not the ENSO itself but the tropical Pacific SST, possibly opening up another avenue of research for this difficult problem.