No need for grep, just use sed's addressing capability

worddiv

@@ -58,20 +58,24 @@ until test "$expression" = "q" ; do
 
        # american
        babel_output=$(cd $tmpdir && latex -draftmode $tmpdir/$expression_filename |  
-			  grep "\\[\\] .OT1" |  sed -E "s/.* ([-a-z]*)$/\1/")
+			      # print nothing (sed -n) except the line that matches
+			      # the regexp "/] .OT1/", but only after replacing the
+			      # whole of that line with part of itself, namely last
+			      # glob of letters and hyphens that comes after a space:
+			      sed -nE "/] .OT1/s/.* ([-a-zA-Z]*)$/\1/p")
 
        # british
        babel_output_secondary=$(cd $tmpdir && latex -draftmode $tmpdir/$secondary_filename | 
-				  grep "\\[\\] .OT1" | sed -E "s/.* ([-a-z]*)$/\1/")
+				  sed -nE "/] .OT1/s/.* ([-a-zA-Z]*)$/\1/p")
 
        ## tex+showhyphens
        tex_output=$(echo "\showhyphens{$expression}" | 
 			tex -output-directory "$tmpdir" | 
-			sed -nE --posix "/tenrm/s/.* ([-a-z])/\1/p")
+			sed -nE --posix "/tenrm/s/.* ([-a-zA-Z])/\1/p")
        
        ## run comparisons ...
        # 
-       # test these by querying the words 'multifacted', 'important', and 'rearranged'
+       # test these by querying the words 'multifacted', 'important', 'rearranged', and 'Switzerland'
        #
        # babel AE vs plain TeX
        if test "$babel_output" = "$tex_output" ; then