In my field (electrical engineering) it is a commonly-accepted practice for publications to be produced at various stages of the "evolution" of a research project, meaning that preliminary results can be published in a workshop paper, intermediate results in a conference paper on the same topic, and "final" or "definitive" results in a journal article. This paradigm is supported by the IEEE, for instance. As long as you cite the original work(s), state that the work is an expanded version of an earlier work (published at the lower stage of the "evolution") and obtain copyright clearances, this is a generally accepted practice.

This leads to a significant reuse of content across generations. I have heard multiple times that the threshold for jumping from a conference to journal paper is "30-40% new material".

My question is this-- do you need to wait for a conference paper to be accepted/published to submit the journal version of an article? If you do not wait, does this violate the "multiple submissions" principle, even if the submissions are not identical, and not at the same "hierarchy". If so, why does waiting for the conference paper to be accepted make this "okay"?

The reason for this question is the case where you are on a tight timetable to submit a journal article, but you want to include some conference results that are currently in review (and may be for several more months).