... repost here in case it's useful:
>
& ... an hour, rather than last week when I was blindly trying ... #3 **A→B→C**, #4 **A←s→B→C**, #5 **A‑m ... #3 is _A→C_, #4 is _A←s→C_, and #5 is ... the conclusion must be _B←s→C_ because **both** variables clearly ...