I sense impatience or even desperation in your original question and comments: If you have got serious indications of you are about to be scooped by other scientists, then I will advise you to publish now on ArXiv or likewise.
In some other scenarios, you could benefit most of keeping your paper secret forever. More on this later on.
Do you realize that your research from a mathematician's point of view most likely is applied science, but might be basic science in the field of natural language analysis? Scientists with a pure linguistic background might not even be aware of the potential of advanced mathematic theories.
I can hardly imagine a single paper suitable for journals in different fields without heavy editing aimed for each separate field: Mathematicians want to read about implementation details of the theory, and why one method works better than other ones. Too less of this stuff and your work is “too soft”. However, they do not care about the implications of your results in the field of linguistics.
Language researchers prefer to know what results you can achieve using advanced mathematics but too many implementation details will look like unnecessary "noise", which never will pass the assessment committee.
Even you do not like the journal's comments I will encourage you to consider that they provides you with useful facts.
Suppose a journal accepted your paper just to make you happy. Language researchers scanning literature for a method to solving a specific problem identical to yours setup, say extract meaning of natural language texts in the style of IBM’s Watson, will not get the point in relation to their fields, because they do not understand larger parts of your paper. Mathematicians looking for examples of appliances of
the theory you applied, say bayesian networks, will quickly move on to papers they can read with lesser efforts.
Only people with experience from both fields can take full advantage of the paper, which implies a very low number of citations. That is not what you want.
Interdisciplinary research often leads to spectacular new knowledge and if your paper contains revolutionary concepts or eye-catching implications, it could be a candidate for Nature or Science without changes. Also, try reading “call for papers” for conferences. If you match their hot topics, they will probably accept your paper as it is.
Otherwise, it is a matter of finding that journal, which you most easily can make your paper fit into, and which is the best choice for your career:
Did you ever ask yourself why you wrote this paper? Academic ambitions? Then I guess only journals of linguistics will count. Do you look for a position as researcher in a private company? Then you should go for applied mathematics.
Based on the comment "definitely scienficially worthy, but too soft" and the professors' general acceptance, it sounds like that the right way for just proper academic recognition could be to rewrite your paper as a book. Wiley, Prentice Hall, etc. tend to listen to recommendations from professors - their actual customers. A submission to ArXiv does not prevent this, and you do not write a single line until you have got a signed contract.
Follow up on just submitting to ArXiv:
A scientist uncovering new areas should always discuss the potential for patents with attorneys before attempting to publish. Publishing prevents patenting. Patenting will worst case only delay publishing.
If your research is useful for software making Siri looking like a toy, I assumes you would like to get along with financial strong companies (Google/Microsoft), which usually keeps their best secrets as secrets and that means no patents/publishing. You deserves economic recognition too.