My awk is a bit rusty so I used sed, which I seem to use almost every day, wonderful, ugly command it is. Here is what I came up with:
dict -d moby-thesaurus back | sed -z 's/,/\n/g' | sed 's/^[[:space:]]*//' | sed -n '/^[dD]/p'
dict looks up "back" in "moby-thesaurus"
sed gulps it all down as a single line using the -z option and, for all commas, substitutes a newline
sed removes all whitespace characters from start of each line
sed prints only lines that start with "d" or "D"
The choice of pattern is important, for example '\<d' also finds multiple word results where one of the words starts with "d" (e.g. "lay down").
But there must be a simpler way. Jeez!
I can hardly wait for general purpose AI.
(Crossposted from https://miriam-e.dreamwidth.org/332278.html at my Dreamwidth account. Number of comments there so far: )