I'm still mucking around with Python, but I did install it, read up on it, installed the Natural Language Toolkit, and a bunch of packages.
Meanwhile, I give you the 100 most common word-clusters/Noun-Phrases from all the prose stories that HPL wrote hisself. It looks like the most common non-stop words (that is, not included in the most frequently occuring English words) are: Time, Old, House, and Night. But Never and Strange are fairly close behind (time and old are far away more frequent). These were collected by software called KHCoder 3, and I got the corpus from GothicChic's pdf, converted to text. I will scrub the stopwords, maybe, but then all the fun stats change and you never know what you'll get that way.
WHO KNOWS WHAT THE FUTURE MIGHT BRING? THE STARS ARE NEARLY RIGHT