Your statement seems to imply that imperative programming does not have modularity. In my experience, I have worked on programs written before OOP was in vogue, in imperative languages, that I consider to have had good modular design. These programs had millions of lines of code and I felt they were easy to understand and modify.

What do you mean when you say that imperative programs lacked modularity?