You are starting from an existing type of (Frankenstein) speaker, and throwing out the challenge to show that it is OK to produce the arbitrary effects that it does, rather than starting at a definition of what a speaker should do and sticking to it. Show me *why* a speaker should spray anti-phase out of the back rather than building such a speaker and retrospectively trying to justify it.
I'm going to hope that you are genuinely curious, that your skepticism is accompanied by open-mindedness (this being a "science" forum rather than a "dogma" forum), and that I have done a really poor job of explaining.
First, my definition of "what a speaker should do": Imo the job of the speaker system is to get both the first-arrival sound and the reverberant sound as perceptually correct as possible.
(If you would like a more philosophical definition of what a speaker should do: Imo the goal is to create the perception of hearing live music, or as close as possible [even if the recording never actually existed as a live performance].)
I'm not going into detail about everything that goes into getting the first-arrival sound right. But let us note that, relevant to the topic of this thread, we get our primary localization cues from the first-arrival sound, after which the Precedence Effect kicks in, suppressing directional cues from the ensuing reflections.
What matters in the reverberant field is sometimes counter-intuitive, so I'll go into a bit more detail here. Most of the sound that most people hear in most rooms is reverberant sound, and most of that reverberant sound started out as off-axis energy. While the reflections contribute relatively little to image localization, they contribute to loudness, timbre, and spaciousness. Done right they can enhance clarity (by giving the ear multiple "looks" at complex sounds, according to Toole), and done wrong they can degrade clarity.
As a general principle, we want the reverberant energy's spectral balance to track that of the first-arrival sound (some reduction in the high end is okay). The reason is, if the reverberant energy is over-emphasizing some part of the spectrum, it makes that part of the spectrum sound louder than it otherwise would have. This skews the perceived tonal balance accordingly and can degrade timbre and clarity and contribute to listening fatigue.
The arrival time of the reflections also matters. Early reflections are generally detrimental to clarity (according to Geddes and Griesinger), though early sidewall reflections can broaden the stereo image a bit, which many listeners enjoy (Toole). In general the earlier the reflections arrive, the more detrimental they are to clarity (Griesinger). In general reflections that arrive roughly 10 milliseconds or more behind the first-arrival sound (that figure comes from Geddes) are beneficial to timbre and spaciousness without any significant downside.
So to sum up:
1. We want all reflections to be spectrally correct regardless of whether they are early or late.
2. We want to minimize early reflections (I'm siding with Geddes and Griesinger, rather than with Toole, on this).
3. We want to encourage beneficial late reflections.
In most rooms, a wide-pattern monopole speaker will generate a lot of early reflections (bad), and a lot of late reflections (good). The spectral balance of the reverberant energy is usually skewed as the radiation pattern usually varies significantly with frequency (bad).
In most rooms, a narrow-pattern monopole will generate fewer early reflections (good), but will correspondingly have less energy arriving as late reflections (not so good, but imo still better overall than a wide-pattern monopole). The spectral balance of the reverberant energy is usually pretty good.
Positioned well out into the room, good dipole speakers generate relatively few early reflections (good) and a lot of late reflections (also good), due to the backwave energy arriving "late". The spectral balance of the reverberant energy is usually good, as that backwave energy is full-spectrum.
As shown by researcher James M. Kates, dipoles generally have smoother in-room bass than monopoles.
When set up properly, good dipole speakers interact with rooms in a way that is arguably more psychoacoustically beneficial than most monopole speakers.
Dipole speakers often have issues in efficiency, bandwidth, bass impact, sweet spot size, power handling, size, cost, and real estate requirements which make them impractical for many people. Just like virtually everything else in speaker design, there are tradeoffs involved.
Does this answer your question?