With reference to [1], the script posted by dokan fails via a Bash syntax error because my version of pdftk 1.41 has multiple lines matching to "Num", which are stored in the TOTAL variable. If we modify the first non-ignored line of the script from
# Original
TOTAL=$( pdftk "$1" dump_data output | grep -i Num | cut -d \ -f2 )
to
# New
TOTAL=$( pdftk "$1" dump_data output | grep -i Number | cut -d \ -f2 )
the script works as expected.
However, its output isn't perfect. In inputting an eighteen page pdf and printing the output, I see that the middle two pages of the document (printed last) are not on the front-and-back of the last sheet, but are instead side-by-side on the same side of the sheet. Therefore, upon converting it to booklet form, instead of having an extra blank "sheet" (half-sheet) at the end of the document, I have two empty pages in the middle of it.
If this (minor but important) issue were fixed, this would be equivalent to printing "Booklet mode" from Acrobat Reader.
Thanks to dokan.