In the text
"The SI unit for momentum is kg·m/s."
Do abbreviations become variables in an expression when they are members of the expression as opposed to simply being a descriptor of the members measurement type? That said, would the kg remain unspaced from the multiplication dot. Would the ; be omitted from m and s. 5.1.2 and 5.18.1 are less than clear for this particular case.
The Nemeth Code does not address this issue directly, nor does it discuss SI units or derived units such as your example, kg·m/s. As you observe, when a unit of measure has no value it is no longer functioning as an abbreviation. Calling it a variable, however, is a bit misleading. "kg·m/s" should be unspaced, but the Nemeth Code requires the use of the English-letter indicator for the "m" and "s", as illustrated in example (8) of Section 51.b of the Nemeth Code.
If this is a chemistry transcription, note that the Chemistry Code does briefly address the issue of SI units in Section 9.3. It shows an example with a multiplication dot, but unfortunately that example has a related value, so the spacing rules of the Nemeth Code are being followed.
Your observation is an important one, and its treatment in Nemeth braille is currently under discussion. Thank you for your question.