What is the explicit meaning of “corresponding author”?
IEEE distinguishes between contact author and corresponding author:
Contact Author: This is the person who is contacted in regards to the submission during the peer-review and production processes.
Corresponding Author: This is the person who is listed as the author to be contacted in the printed publication.
PLoS states two types of corresponding authors:
Please note that the corresponding author for the online submission process can be different from the corresponding author of your published manuscript.
Copernicus uses the terms contact author and corresponding author:
From submission to publication, the authors of a manuscript are formally represented by the registered contact author(s). After publication, the authors of a manuscript are formally represented by the corresponding author(s) specified in the published paper.
Personally, I've seen this differentiation several times. Sometimes the corresponding author is simply a role in the journal system that the submitting author/person can assign to one of the authors during the submission process. The corresponding author might also change during the process.
Conclusion:
Some publishers make a difference between the author corresponding with the journal (submitting/contact author) and the author corresponding with potential readers of the article (corresponding author). Some publishers don't. It's absolutely inhomogeneous and seems also to vary by journal not only by publisher. And it varies over time.
Side note:
And yes, these publishers are indexed in SCOPUS and Web of Science (Science Citation Index, Thomson Reuters Master Journal List).
Search for the publishers in SCOPUS (select checkbox "publisher") or the Thomson Reuters Master Journal List (TRMJL). The latter includes all journals indexed in Web of Science. Unfortunately, the search in TRMJL is only based on journals and not on publishers. Search for "plos" to see PLOS journals, "ieee" for IEEE journals and e.g. "geoscientific" to see two Copernicus journals. See also IEEE indexing agreements, Visibility of PLoS and Copernicus metrics.
I don't think you're going to find anyone who has the final definition for the term. It is not defined by the law or some government regulation, but instead is a term that is used by speakers in the community who may have different interpretations and nuances on their minds when they use it. As a consequence, I suspect that all you can find is how people interpret the term, but you won't get universal consensus. In the current context, it also seems to me that not very much is lost if there is no universal consensus.
To me, and I suspect to most in the community, the "corresponding author" handles both roles: communication throughout the publishing process, and about possible questions. In the past, the corresponding author may also have been the one who has off-prints of the article sitting in his office that a potential reader could request (by mail), but this is no longer happening in times of the internet. Rather, if you have questions, you may contact any of the authors of a paper.
While maybe not a full answer across the fields, in some journals, the papers have a description about what it means.
For example, in IOP papers (a paper of mine where it does happen) it says:
"Author to whom any correspondence should be addressed."
Which, considering the paper has already been published, means the correspondence about any further questions one may have about the research/science on the paper. Note, I did not submit the paper.
However, to give a counter example, IEEE guidelines describe Corresponding author as (sec 6, page 11):
The corresponding author is responsible for submitting the manuscript and managing it through the review and revision process with the publisher. The corresponding author makes sure that all authors are kept apprised of the current status of the work.
Which could be considered the "submitting author" as in the question linked there.
I guess that the answer is: The definition is not written in stone and different editors/journals use it differently