@2184a19c Oh, I'm happy to share it!  Thinking of writing it up as a blog post, because I think it deserves more than to be buried in a discussion of propagation and Croft's lame extended "Lewis and Clark" joke. 

 @2184a19c That would be William Croft (2000). Explaining Language Change: An Evolutionary Approach.

I could have sworn there was a whole big discussion of it, but it seems to be buried in section 7.4 (pages 174-194) on propagation.  Seemed like a huge deal to me when I read it!

(full disclosure: Bill Croft was on my dissertation committee, and my feelings about him are ... complicated)

https://books.google.com/books?id=5_Ka7zLl9HQC 

 @2184a19c The City of Lost Children! 

 @2184a19c Here's the relevant page; on the previous page and a half there is a discussion of "grammaticization" vs. "grammaticalization," which had no sense distinction anyone could find, and was therefore interpreted socially or geographically.

"there seems to be a natural human tendency to increase the conventionality of one variant of a lingueme in a community at the expense of another, albeit over a long period of time in many cases. We may call this the FIRST LAW OF PROPAGATION."

https://cdn.masto.host/lingolol/media_attachments/files/111/182/817/718/729/567/original/bfd1caeaa56181f6.jpg 

 @2184a19c Croft has a whole section in his 2000 book about language abhoring pure synonymy - if there's no semantic, pragmatic, situational or social distinction, people will invent one!