This article examines a way to define left and right versions of the large class of “ ( b , c ) -inverses” introduced by the writer in (2012) [6] : Given any semigroup S and any a , b , c ∈ S , then a is called left ( b , c ) - invertible if b ∈ S c a b , and x ∈ S is called a left ( b , c ) - inverse of a if x ∈ S c and x a b = b , and dually c ∈ c a b S , z ∈ S b and c a z = z for right ( b , c ) -inverses z of a . It is shown that left and right ( b , c ) -invertibility of a together imply ( b , c ) -invertibility, in which case every left ( b , c ) -inverse of a is also a right ( b , c ) -inverse, and conversely, and then all left or right ( b , c ) -inverses of a coincide. When b = c (e.g. for the Moore-Penrose inverse or for the pseudo-inverse of the author) left ( b , b ) -invertibility coincides with right ( b , b ) -invertibility in every strongly π -regular semigroup. A fundamental result of Vaserstein and Goodearl, which guarantees the left-right symmetry of Bass's property of stable range 1, is extended from two-sided inverses to left or right inverses, and, for central b , to left or right ( b , b ) -inverses. [ABSTRACT FROM AUTHOR]