sS ai_groupTheAdjectiveNoun(S s) {
  ret join(groupTokenPredicateSequence(wordTok(s), ll(curry eqic("the"), f isAdjective, f isNoun));
}