diff --git a/scripts/textometry/compute-profile.hs b/scripts/textometry/compute-profile.hs
index 4ac3480f2683af28120cd433ca071c9f00c05296..04aa4e2c818a4e5bd577f5d7a34f6c1c0ebe938d 100755
--- a/scripts/textometry/compute-profile.hs
+++ b/scripts/textometry/compute-profile.hs
@@ -2,11 +2,12 @@
 
 {-# LANGUAGE DeriveGeneric, ExplicitNamespaces #-}
 import Conllu.Tree
-  ( Feat(..), IndexedDocument(..), IndexedSentence(..), IndexedWord(..)
+  ( Feat(..), IndexedDocument(..), IndexedSentence(..), IndexedToken(..)
   , positions )
 import Control.Applicative ((<**>))
 import Control.Monad.IO.Class (liftIO)
 import Control.Monad.Reader (ReaderT, asks, runReaderT)
+import Conllu.Tree.Count (Count(..))
 import Data.Csv (DefaultOrdered(..), ToNamedRecord(..))
 import Data.ByteString as ByteString (readFile)
 import Data.List (sort)
@@ -63,11 +64,13 @@ profile articleRecord = do
     analyse = pure . fmap (glue articleRecord) . findOccurrences
 
 findOccurrences :: IndexedDocument -> [Occurrence]
-findOccurrences (IndexedDocument {_total, _sentences}) =
+findOccurrences (IndexedDocument {_total = Count {tokens}, _sentences}) =
   groupOccurrences [] (_sentences >>= imperativeVerb)
   where
+    occurrence (position, size) =
+      Occurrence {event = "Imperative", position, size, totalSize = tokens}
     groupOccurrences :: [(Int, Int)] -> [Int] -> [Occurrence]
-    groupOccurrences stack [] = uncurry (Occurrence _total) <$> reverse stack
+    groupOccurrences stack [] = occurrence <$> reverse stack
     groupOccurrences [] (p:ps) = groupOccurrences [(p, 1)] ps
     groupOccurrences (tmp@(first, tmpLength):stack) (p:ps)
       | p == first + tmpLength =