Alex, I wish you well and look forward to reading your articles on "Truth Units."
At a surface grammar level "Truth Units" is a predication of "Units" modified by "Truth" but it is a form of two nouns so it can simultaneously also be the opposite predication of "Truth" modified by "Units". So it should be possible to say Truth Units and Unit Truths and mean the same thing. Which you can, I think. But human natural language is a horrible place to be truthful in.
In my work on lies there are two definitions I use for lies One is the common legal definition which is purely set theoretic and uses only the opposite of "lie" which is "truth" in the definition: "the truth, the whole truth, and nothing but the truth." Anything that isn't the truth in this complete set theoretical sense is a lie.
For the second definition, there is a taxonomy of all lies and truths. At the top of the taxonomy is whether an assertion is a suspected, or fiat, lie, considered to be certainly a lie, or considered to be true by someone else. A fiat lie is a suspect lie for which dialogue about it as a conversational lie is not possible for any reason.
The taxonomy of conversational lies that you can dialogue with another person comes out of what are linguistically the presuppositions of any natural language use of the linguistically universal word-concept 'lie'. This includes the answers to the questions of (1) is it the truth? (2) the whole truth? (3) nothing but the truth?(4) what is the deceit? (5) was the deceit intended? (6) what is the motivation for the lie? (7) is the lie socially acceptable? and (8) in what contextual framing (i.e. further linguistic predication) can it be made more socially acceptable?
These are the universal questions you can ask of any lie.
That said, the computational cognitive neuroscience of lies studies what anyone can call a lie and be accepted by someone else as being a true assertion about what is being called, predicated, as a 'lie'.
My book on "How to get your (good) lies back" where I develop how the brains compute on lies, I mention that it is same mechanism they use to compute on their own privacy in communications. This is volume one, another book, "How to get your privacy back". And there I mention the third volume which I chose not to write, as an exercise for the reader, on "How to get your truth back".
Here is a short summary of the science of lies and truth too from a computational cognitive neuroscience context or "very Cartesian" context:
Here is how any brain most fundamentally computes and how human natural language, asserted truth and lies, is computed out of its computations.
https://medium.com/liecatcher/natural-language-and-your-brain-237185770b00
So, truth is always relative to its predicative context which is what brain computation is all about. Any truth can be a lie, just as any lie can be a truth if all you have is the natural language expression or the natural language understanding.
That said an artificial language can always be constructed so disagreement as to truth or lie can be universally accepted. This is what happens, for example, in Physics with the rules of measurement and math.
https://medium.com/liecatcher/lies-and-truth-come-from-the-same-place-b7dc2fb212bc
I've been studying this computational cognitive neuroscience since 1970ish. Published my first paper in 1973 and my dissertation on it in 1976.
That said if you can figure out what a truth unit is in a way that makes sense outside of a primitive linguistic predication, I look forward that insight. I look forward to your articles on your ideas.
