Except it's not generated. A simple google search [1] reveals there's at least one paper out there that contains a structurally identical phrase [2]. The "Dada Engine" just takes real papers and substitutes nouns and pronouns. A similar strategy could probably be used on most any academic paper. But it does appeal to those who would never attempt to read these papers in the first place.
> The "Dada Engine" just takes real papers and substitutes nouns and pronouns. A similar strategy could probably be used on most any academic paper.
I dare you to do this on e.g. a Physics paper. This doesn't mean badly written Physics papers don't exist, but unless it's an awful paper or you aren't familiar with that field, the statistics and data analysis should stand on it's own. So you always have this fallback. Of course they aren't infallible either, but it's one extra BS detection layer.
For social science, extra BS detection could be had via reference checking. Unfortunately, this is effort and tedious, in part because almost all papers are long-winded and hard to read. So instead, form/style/reputation of cited authors/buzz-words is used. All of these are easy to fake.
I don't know what the solution is. I've seen in Philosophy, students would mark text with logic symbols. Then, you can see a bit more easily if "therefore" really means something follows. Maybe that kind of rigorousness is needed, in an explicit form?
But you're right in that it's mostly laziness, and the random generators don't hold up to too much scrutiny. It's more that it's a mirror, a way of pointing out the shortcomings of the field. Which is why it's funny.
I dare you to do this on e.g. a Physics paper. This doesn't mean badly written Physics papers don't exist, but unless it's an awful paper or you aren't familiar with that field, the statistics and data analysis should stand on it's own. So you always have this fallback. Of course they aren't infallible either, but it's one extra BS detection layer.
For social science, extra BS detection could be had via reference checking. Unfortunately, this is effort and tedious, in part because almost all papers are long-winded and hard to read. So instead, form/style/reputation of cited authors/buzz-words is used. All of these are easy to fake.
I don't know what the solution is. I've seen in Philosophy, students would mark text with logic symbols. Then, you can see a bit more easily if "therefore" really means something follows. Maybe that kind of rigorousness is needed, in an explicit form?
But you're right in that it's mostly laziness, and the random generators don't hold up to too much scrutiny. It's more that it's a mirror, a way of pointing out the shortcomings of the field. Which is why it's funny.