I'm glad you only provided the three choices. Whether we are aware or not, the community makes hard choices every day which take from one of these three disciplines and give to one or more of the others. There's talk of convergence and evolution which would eliminate the barrier between these discreet disciplines. While certainly possible, it is by no means sufficiently realized or ubiquitous to utilize as the philosophical underpinning for future plans. Yet, we have demonstrated a proclivity over the past decade to project our philosophy (see Information Warfare community, the FIOC CONOPs, NUCO, distributed operations and other abortions) in a way which doesn't directly address the priorities of the Commanders we are supposed to support. Tough to figure if we are or are not wrapping future reality around what we'd like to see rather than what is happening. The point is - we can't hold hostage making tough choices/practical decision making because we'd like to believe one or more disciplines can't exist without the other. The fact is, while they are very complementary, each discipline can - and does - exist separately from the others. Considering the types of adversaries we face and the universally understood risk of exploitation inherent across the entire EM, it makes sense that each discipline will continue to exist separately.
Separately, I'm beginning to think we've reached the point where cyber and SIGINT/EW must split into separate efforts/communities. Given how we are structured/our numbers/$/etc, I don't see how we can take the necessary steps in either area to efficiently move forward. We do have occasion where we MUST be able to decide between the two and that makes many very uncomfortable, even though it shouldn't. There is no sacred cow and no sacred process which must include or conflate disciplines together (other than maybe TOA, flag poles, etc). Cyber is vital for cyber mission teams - SIGINT/EW, not at all. SIGINT/EW is vital for ships; cyber (1810/CT type cyber - not to be confused with DODIN OPs), not at all. Understanding where authorities lie, CYBERCOM (as a possible full COCOM), and each discipline: there are far more reasons to support intelligent division of what is now an excessively large and overly general job jar. That specialization (cyber and SIGINT/EW), latitude, and focus would allow each resulting community to make better headway independently than our one combined community is capable of making today.