It is indeed interesting but, for me, not convincing. The differences that Tobin claims may be present for some speakers/writers, but my (not very scientific) observation over the years suggests that:
1. For most speakers and writers they are exact synonyms.
2. Many speakers and writers tend to use mainly one of the two forms.
3. Where a difference is felt, it is generally that 'till' is less formal.
The dictionaries and grammars I have consulted agree with me on points #1 and #3.
If it be true that most speakers do not feel the difference that Tobin notes, then it is not a real difference , except for a minority of speakers/writers. And, even if it true for some, this becomes irrelevant if the majority of listeners/readers do not appreciate the subtle point being made.
Personally, I use whilst and while with subtly different meanings, but most people don't, so I could save myself the trouble.
If you analysed my speech and writing, you would find that I use may and might with subtly different meanings. I am not alone in this and, historically, they did have different meanings, might being the past-tense form of may. However, most speakers and writers appear not to feel a difference today; indeed many Americans do not use may at all. There is no point in those who of us feel a difference preserving this. (That will not stop me doing it, but that is irrelevant.)
So, as I suggested above, even if some speakers and writers do use till and until to suggest different meanings, most of us are not aware of it,and it is therefore not relevant to learners.