There are a lot of variations in the original ContextD triggers, which could be caught with more general patterns.

• There are a lot of variants of geen (e.g. geen aanwijzingen voor, geen last van), but there is also geen itself, which would already match all of its variants...
• We could match on the lemma of a token, such that e.g.
  • "geen teken van", "geen tekenen van", "geen tekens van" will be caught in one rule
  • "vertoonde geen", "vertoont geen" will be caught in one rule
• There's a lot of variants of "als de/een [bron/oorzaak/reden] van/voor" which could be caught in one rule (e.g. by optionally matching any token)