I'm very familiar with the literature on sensory substitution. However,  I am suggesting that the body already has a hard-wired sensory substitution device (i.e. that there is a previously unexplored relationship  between vision and proprioception/motor imagery). If this built-in sensory substitution capacity did exist, it would have an extteme advantage over artificial devices, because it would not require as much training; the brain would already be hard-wired to substitute information from one modality to another. That's one of the reasons sensory substitution devices like the VOICE are rarely used outside the laboratory; they require a pretty significant learning curve. I am also suggesting that for some blind people, (me very much included), spatial processing is impaired, not only because vision is absent, but because retinatopic-like maps failed to develop in other senses (I'm mainly talking about hearing). My friend Lindsay, who is blind due to Leber's Congenital Amaurosis, just like I am (though she might have a different mutation), is very easily able to generate mental pictures of space, based on sound and touch. I, on the other hand, very much lack this ability. However, smoking marijuana not only gives me access to some vision (see my paper), but also an extremely strong sense of verticals, horizontals, surfaces and edges. This information serves as a grammar or template into which I can plug in information from multiple senses. What I am trying to get at here is that if you are blind, and that "visual grammar" doesn't develop, you are doubly disadvantaged, and that if sound or touch fail to help you develop that grammar, proprioception might be another viable candidate. My idea of a visual grammar is taken from  Patrick Cavanagh's work.

I think...there should be a huge body of work from the field of visual impairment, and sensory deprivation. When there is no illumination to process a 'sense' of space and self remains.

There is already significant research on this topic, known as sensory substitution. It's a really interesting research field.

As you've predicted, individuals can map information from one sensory modality and represent it in another modality. In other words, people can learn to 'see' using their skin (although the definition of seeing might be loose in this context, though see this paper: Striem-Amit, E., & Amedi, A. (2014) for an argument that participants truly are seeing). 

Bach-y-Rita pioneered this research in the 1960's. You can find some relevant works by searching for Sensory Substitution on google scholar.

Typical experiments present information to one sensory modality (e.g. vibrating units placed on a person's back) and correlate the input with some physical dimensions in space (which we typically decode visually, as you've considered). These experiments have been conducted with all kinds of sensory pairs, including sound to sight, touch to sight, and I'm sure more are in the works. 

Note that the cross-sensory mapping isn't retinotopic, as you've suggested. The mappings from one sensory modality to another can be quite flexible. This means that up (in space) might be up (on the skin), but it could just as easily map in the reverse direction. This is particularly true for substitution devices that go onto areas that move a lot in space, or don't have a clear up/down orientation (e.g. the tongue or the palm of the hand). 

Some interesting solutions for optimal mappings emerge when participants are allowed to adapt the Sensory Substitution Devices to their own preference using genetic algorithms (see Wright & Ward, 2013) .  

Here are some relevant references:

Bach-y-Rita, P. (Ed.). (1972). Brain mechanisms in sensory substitution. Academic Press.

Striem-Amit, E., & Amedi, A. (2014). Visual cortex extrastriate body-selective area activation in congenitally blind people “seeing” by using sounds. Current Biology, 24(6), 687-692.

White, B. W., Saunders, F. A., Scadden, L., Bach-Y-Rita, P., & Collins, C. C. (1970). Seeing with the skin. Perception & Psychophysics, 7(1), 23-27.

Wright, T., & Ward, J. (2013). The evolution of a visual-to-auditory sensory substitution device using interactive genetic algorithms. The Quarterly Journal of Experimental Psychology, 66(8), 1620-1638.

https://scholar.google.com.au/scholar?hl=en&q=sensory+substitution&btnG=&as_sdt=1%2C5&as_sdtp=

I'm very familiar with the literature on sensory substitution. However,  I am suggesting that the body already has a hard-wired sensory substitution device (i.e. that there is a previously unexplored relationship  between vision and proprioception/motor imagery). If this built-in sensory substitution capacity did exist, it would have an extteme advantage over artificial devices, because it would not require as much training; the brain would already be hard-wired to substitute information from one modality to another. That's one of the reasons sensory substitution devices like the VOICE are rarely used outside the laboratory; they require a pretty significant learning curve. I am also suggesting that for some blind people, (me very much included), spatial processing is impaired, not only because vision is absent, but because retinatopic-like maps failed to develop in other senses (I'm mainly talking about hearing). My friend Lindsay, who is blind due to Leber's Congenital Amaurosis, just like I am (though she might have a different mutation), is very easily able to generate mental pictures of space, based on sound and touch. I, on the other hand, very much lack this ability. However, smoking marijuana not only gives me access to some vision (see my paper), but also an extremely strong sense of verticals, horizontals, surfaces and edges. This information serves as a grammar or template into which I can plug in information from multiple senses. What I am trying to get at here is that if you are blind, and that "visual grammar" doesn't develop, you are doubly disadvantaged, and that if sound or touch fail to help you develop that grammar, proprioception might be another viable candidate. My idea of a visual grammar is taken from  Patrick Cavanagh's work.

Now that I've read more about predictive processing and the role of proprioception in motor planning, (thanks Link!) I have a couple more ideas as to what could be going on. It's possible that my idea of proprioception allowing the body to serve as a computer screen is just my brain's translation of what's going on. If proprioception functions as motor commands, perhaps the intense  association between proprioception and my visual experiences under marijuana could be my brain's way  of sending me information about how I should best move so that the data from visual prediction error is more in line with priors. Or perhaps the increased weighting of proprioception might be my brain's way of trying to enrich impoverished generative models of visual/spatial processing (i.e. using my body as a model to understand and predict objects located in external space). The problem now is that   there is too much gain on the error signal from proprioception, so I'm conflating self-generated movement with movement in my environment.  I'd like to say that stabilizing  my nystagmus would rectify this, but the tendency for my proprioceptive system to try to "mirror" what I'm seeing is not restricted to the eye muscles.

Dear Thasha, 

Concerning the generative models and the relationships between the real phase spaces (the environment) and the abstract ones (the brain), look at: https://www.researchgate.net/publication/297679077_Towards_a_Neuronal_Gauge_Theory

Article Towards a Neuronal Gauge Theory

Thanks so much for this article. I wish my knowledge of statistics/computational neuroscience was sophisticated enough to fully appreciate it. Would it be possible to give me a simplified explanation of how you theorize the relationship  between physical and abstract  space? I read Andy Clark's Surfing Uncertainty so have a basic (non-technical) understanding of Bayesian brain theory.

dear Tasha, if you Read just the Box 1 at page 8, all things Will become clearer!