Please don't give me the Wikipedia or dictionary explanations, we're looking at mutual influence.
In computer science the primitive expression of syntax would be as the format for instructions to the processor: "move x to w" or "w,x mov"; "add y,w" or "w+y"; etc. The syntax is determined ultimately by processor architecture, we can think of it as the rules that determine whether the computer can 'understand' the semantic (data) content. Given correct syntax, the computer knows what to do with the data but it is the programmer that must process the semantic content into a good syntax. From syntax, the computer can learn nothing, but a programmer can learn about the architecture of the computer.
For natural language, syntax is less important. The semantic content can be successfully transmitted even if the message syntax is imperfect. In addition (extra) semantic content can be encoded in non standard ways by variations in syntactic form; for example: "John is kicking the ball!" also mean he is kicking high, and "The ball is being kicked be John." means he is kicking deep... In addition to pragmatics, semiotics, and other linguistic features, humans can manipulate syntax to convey meaning.
It seems there are more than one level available for the processing of language. There is the literal (computer) level, then there is an adaptive (self-programming), communal (interactive) level of language usage. All of this brings up three questions, any one of which I would appreciate an answer to:
1) Is there similar issues in formal languages, and how are they treated?
2) Are there any literature that explores this in the context of "brain-mind interaction", "consciousness ", "self-awareness" etc?
3) What implications are there for artificial intelligence research, and has there been any publications?
I realize this is rather 'fringe' for Philosophy, overlapping with linguistics, computer and cognitive science, but there is a common thread (I think) that could have strong epistemic implications.