Eggcorn Forum

It sounds perfectly right at first, probably because the word-pair “sum up” is so common in its own right.
Lots of wrong ones on Google, and one using this phrase in a literally correct sense:
http://findarticles.com/p/articles/mi_h … _n18507151

Ditto here, Joe. Hey, by the way, when you make it to 1000 posts, do you earn a new title? ....”Eggcornquistador”?

Interesting—I never really thought about all those optional “ups” in verb forms before. To see how often they were used, I did some googling. Here are the results:

muster the troops 6310 (71.6%)
muster up the troops 2500 (28.4%)

muster the courage 109000 (70.4%)
muster up the courage 45900 (29.6%)

summon the courage 64300 (68.3%)
summon up the courage 29800 (31.7%)

Considering that my raw data comes from Google inquiries, these results are amazingly consistent. The forms with “up” are running about 30%—a very respectable share. I imagine the figures for the “up” forms would drop if I looked at edited prose, but they’re clearly in very wide use. My guess would be that people who tend to use “up” with one of the verbs tend to use them with the other ones, too. And the same pattern applies to “up”-averse speakers.

Of the various verbs here, summon seems most likely to go with “up” – perhaps because of the spiritual sense of summoning up demons, the devil etc – of course they are summoned up from ‘down there’, so the up is literal (as far as a reference to hell can be real or literal)

Welcome back, Gilibug!

I was amazed by Gilibug’s numbers for “rally the courage.” It’s not a phrase you hear much, and 12.5k for a 3-word fixed string on Google is pretty darn respectable. I ran the search myself and got even more—12.6k. But—suspicious—I decided to figure out the unique hits. Guess what? That 12,600 dwindles rather startlingly to 38 hits.

That still gives “rally up the courage” a rather low percentage (about 5 %) against the other “up” phrases. But if Google gets hold of this post, the percentage will rise to about 7%. Maybe I’ll use “rally up” in a few other posts in order to achieve the 30% I’d predict.

Before I get busy manipulating my own data, this is probably a good time to remind people once again of the mounting problem of using raw Google numbers rather than unique Google hits. (I last raised the issue a few days ago in regard to Eggstatic’s post on “pure bread.”) If raw Google numbers seem too big to be true, they almost certainly are.

Last edited by patschwieterman (2007-09-20 23:01:37)

Eggstatic—Yes, one way to get the unique hits for a Google search is simply to scroll to the end of the pages returned.

Of course, if you get much more than 100 hits, that can start becoming tedious. So for searches with lots of legitimate hits, another way to do it is to add the string “&start=950” (without the quotes) to the Google search URL. Let’s take the word “mind-bottling” from a recent post by Faldage as an example. If you put in “mind-bottling” at Google, the end of the search string might look like this:

search?hl=en&q=%22mind-bottling%22&btnG=Google+Search

That search gives you 24,500 hits. But if you add &start=950 to it, the end of the string will look like this:

search?hl=en&q=%22mind-bottling%22&btnG=Google+Search&start=950

That search returns only 489 hits, and some of those aren’t the eggcorn. So that demonstrates the value of looking for unique hits—the raw hit total inflates the real numbers for “mind-bottling” by over 24k.

Here’s my actual google search for “mind-bottling”:

http://www.google.com/search?hl=en&q=%2 … gle+Search

Here’s the same search with the “unique string” added:

http://www.google.com/search?hl=en&q=%2 … &start=950