The lessons of “logical empiricism” still need learning by scientific philosophers who continue to give ultimate descriptions of ultimate things. Even attempts to assign some basic transformative role to language must fail, as this role can always be reversed on some higher language level. Final philosophical characterizations of the universe collapse when we see their arbitrary linguistic nature. Similarly, the effort to force scientific philosophy into an ontology, as shown by a recent example, turns out to be only a conventional way of talking, rather than a legitimate appeal to something in the universe. More vulnerable recent examples of scientific metaphysical descriptions of “reality” reveal a deceptive linguistic superfluity in no way necessary for science or scientific philosophy. Within the inevitable circularity of language, semantic analysis can still fasten to the goal of predictability; and predictability can do without uneconomical symbolism.