Inasmuch as physical theories are formalizable, set theory provides a
framework for theoretical physics. Four speculations about the relevan
ce of set theoretical modeling for physics are presented: the role of
transcendental set theory (i) in chaos theory, (ii) for paradoxical de
compositions of solid three-dimensional objects, (iii) in the theory o
f effective computability (Church-Turing thesis) related to the possib
le ''solution of supertasks,'' and (iv) for weak solutions. Several ap
proaches to set theory mid their advantages and disadvantages for phys
ical applications are discussed. Cantorian ''naive'' (i.e., nonaxiomat
ic) set theory, contructivism, and operationalism. In rite author's op
inion, an attitude of ''suspended attention'' (a term borrowed from ps
ychoanalysis) seems most promising for progress. Physical and set theo
retical entities must be operationalized wherever, possible. Ar the sa
me lime, physicists should be open to ''bizarre'' oi ''mindboggling''
new formalisms, which need not be operationalizable or testable at the
time of their creation, but which may successfully lend to novel fiel
ds of phenomenology and technology.