On a supposed advantage of realistic systems

In The Structure of Appearance' Nelson Goodman at one point considers the respective merits of two kinds of systems which he calls 'particularistic' and 'realistic'. A particularistic system is one where 'concrete spatially or temporally bounded particulars (such as phenomenal events) are taken as basic units', a realistic system one where 'nonconcrete qualitative elements (such as qualia)' are taken as basic units.2 An example of a particularistic system is the one presented in Carnap's Der Logische Aufbau der Welt;3 an example of a realistic system is Goodman's own, presented in Chapters 6-11 of SA. Although proponents of realism and particularism sometimes offer arguments designed to show, in advance of any constructions, that their kind of system is more likely to be ultimately satisfactory than the other kinds, Goodman is inclined to be tolerant: he counsels patience, pointing to the value of having systems of various sorts developed, given what we learn from the exercise. But he does point out, in his exposition of the Aufbau, that Carnap's system runs afoul of two kinds of difficulty which he christens the 'companionship difficulty' and the 'difficulty of imperfect community.'4 Later,5 he shows how adoption of a nominalistic framework based on the calculus of individuals will save a particularistic system from the difficulty of imperfect community, but not from the companionship difficulty. But, he argues, a realistic, nominalistic system does avoid the companionship difficulty, and so,