walkman - 1:38 am on Feb 28, 2011 (gmt 0)
My content is all original, it's well-researched, but, since 85% of the pages on my site are database driven "entries" almost like a dictionary, it's quite possible that the overall look, to a machine, is one of a bunch of "similar" pages. I don't think that's the whole story, but I think Whoa's comments ring very true to me. They point in the right direction, I think. (Or at least one of the right directions!).
That may be part of the story, but most 'content farms' had decent size articles and definitely not stubs, nor can you say that their sentences aren't complete. Stupid and useless content can be very good grammatically.
We also have had this debate for ages: at what point does the google's duplicate penalty kicks in, and talked about too many tags (brings same short stubs), large footers etc.