Catching criminals and watching them in the act

CW: mention of public indecency behaviour in the tweet linked and below, before we get to the linguistics. 

This reply to a tweet of Derren Brown's delighted me. The response is to a screenshot of a takeaway driver review in which the person says they 'caught the delivery guy playing with himself in his car for ten minutes': 

Ten minutes seems a long time to "catch" someone doing something. I feel like that strays into "watch".

— Alan ‘join a union’ Smith (@mralanjohnsmith) December 19, 2021

As Alan says above, you can't really catch someone doing something for ten minutes. Why not? 

This is what linguists call 'lexical aspect', and it's basically part of the meaning of the verb, but it also interacts with the tense and other features of the way you use it. It's why if you say I'll go for a walk in an hour you mean that an hour from now, you will go for a walk, whereas I finished my homework in an hour means that it took you one hour to do your homework, and I'll do my homework in an hour is ambiguous: you'll do it in one hour's time, or it will take you one hour to finish it. Notice that I'll go for a walk in an hour doesn't have this ambiguity because it's just not the type of verb that you can say how long it takes to complete. You can, however, say I'll go for a walk for an hour, and then you're expressing the duration of your walk. 

There are a few different systems of classifying verbs in this way, but let's go with the classic one: Vendler's system from 1957. On this classification, catch is what he termed an 'achievement': an instantaneous action, a point in time. While you might have a long build-up to the catching, like a year-long investigation and stake-out, the act of catching itself is an instant in time. Watch, on the other hand, is an 'activity', which is an ongoing process without a pre-defined end point. If it doesn't have a pre-defined end point, then you can specify how long it went on, as in this case (ten minutes). 

You can manipulate these classes, and for instance say that It took ten years to catch the criminal, and then although we're using a frame that specifies the length of the process, that length of time is actually the delay before the achievement takes place: ten years of meticulous planning and investigating, or perhaps alternatively bumbling incompetence. This is why, after the ten years is up, you might say We finally caught the criminal on Tuesday, after ten long years!. Compare this with an 'accomplishment' like write a novel: you can say It took ten years to write my novel, but you can't say I finally wrote my novel on Tuesday, after ten long years!. You'd have to say instead that you finished it on Tuesday, because write a novel includes the process leading up to the completion as well as the completion itself, unlike catch, which is just the completion and not the process.