proceedings of the
american mathematical society
Volume 113, Number 1, September 1991
SOME REMARKS ABOUT ROSEN'S FUNCTIONS
ZBIGNIEW GRANDE
(Communicated by Andrew M. Bruckner)
Abstract. The main result is: Each Baire 2 function /:/—»/? whose set of
continuity points is dense is the pointwise limit of a sequence of Darboux Baire
\ functions.
Let / = [0, 1] and R be the set of all reals. A function /:/—>/? is said
to be Baire \ [6] if preimages of open sets are G^-sets. (Rosen states these
functions as Baire -.5 [6].) In [6] H. Rosen shows the following theorem:
Theorem 0. Suppose f: I —► R is a Darboux Baire \ function, and let D denote
the set of points at which f is continuous. Then the graph of f/D is bilaterally
c-dense in the graph of f.
Remark 1. A function f:I—>R is a Baire j function iff it is ambiguously a
Baire 1 function, i.e. preimages of open sets are Gs- and i^-sets simultaneously.
Indeed if / is Baire \ , then every open set U is the sum of closed sets Fn
(n - 1, 2, ...), hence
(oo \ oo
n=l ) n=l
and U~, r\Fn) is an F„-set.
Remark 2. A function /:/->Ä is said to be a.e. continuous [4] if it is ap-
proximately continuous and continuous almost everywhere (in the sense of the
Lebesgue measure). There is an a.e. continuous function f:I—>R which is
not Baire j.
Example 1. Indeed, let P c / be a Cantor set of measure zero and let A c P
be a countable set such that CIA = P (CIA denotes the closure of the set A).
Let (an)n be a sequence of all points of the set A . There is a family of closed
intervals Jnm c I - P (n, m = 1, 2, ...) such that:
(!) JnmnJrs = 0 ü (n, m) ¿ (r, s), n, m, r,s = 1, 2, ... ;
(2) an is a density point of the set (Jm=i Km > n = 1,2, ... ;
Received by the editors March 9, 1990.
1980 Mathematics Subject Classification (1985 Revision). Primary 26A21; Secondary 26A03,
26B05, 28A20.
© 1991 American Mathematical Society
0002-9939/91 $1.00+ $.25 per page
117
License or copyright restrictions may apply to redistribution; see https://www.ams.org/journal-terms-of-use