100 million word British National Corpus.

[Tdol]

In that sense, though it generates error, it is the most accurate picture of English use of all.

"depends of" - 176,000
Yikes!

Originally Posted by tdol

In that sense, though it generates error, it is the most accurate picture of English use of all.

"depends of" - 176,000
Yikes!

It gets worse:

Googled:

155,000 English pages for ABSORBTION.
37,200 English pages for "conflicting feelings".
4,420 English pages for "I me myself ".
1,790 English pages for "married with him
571,000 English pages for "grammer".
759 English pages for "he teached me".

BNC:
BNC:

• ABSORBTION 4
• CONFLICTING FEELINGS 6
• I me myself 0
• got married with her 0
• he teached me 0
• many photo of 0
• office equipments 0
• acquainted to 0
• was borned 0

[Tdol]

There are so many students searching for 'grammer' that many sites include the word to bring in the students through their error.

Originally Posted by tdol

There are so many students searching for 'grammer' that many sites include the word to bring in the students through their error.

Let's hope the porn sites don't catch on to that spelling and tactic then.

[Tdol]

I'm sure they already have for many common mistakes.

[wordhound]

To find the most common words in the media, why not just Google them? I find the corpus opaque, it must be taken on trust, Google is open and transparent.

[wordhound]

Sept 2005, Google Word Hits

the 3.5 billion
to 3.4
and 3.3
of 3.2
a 3.2
for 3.2
in 3.1
on 2.8
home 2.7
about 2.6
site 2.5
is 2.5
all 2.4
from 2.3
search 2.3
at 2.3
you 2.3
this 2.2
web 2.2
our 2.2
more 2.2
new 2.1
your 2.1


more to come

Originally Posted by wordhound

To find the most common words in the media, why not just Google them? I find the corpus opaque, it must be taken on trust, Google is open and transparent.

How would I do that in Google? I mean, how could I separate all other uses from uses in the media. What would my search term be?

[wordhound]

When I took a TEFL/TESL course last year I got the list of the most common words in rank order. The same handout also said the 8 words are 33% of spoken English. I thought later, how do they know without some numbers attached? Most common means most used. So counts can be established but this is never done. What we get are statistical analysis like joint frequency. This corpus is taken on faith, we trust someone to do a good job. Where can we verify?

Most lists get the first eight words right, then after that it varies. How about some consistency? Google is consistent.

I suspect web companies use this system already but the ESL world has declined to notice it.

[Tdol]

I use both Google and the BNC for language concordancing. No matter how big, any database will naturally have a certain degree of innacuracy or bias built in, based on the choice of texts and sources. However, the BNC, at sites like http://view.byu.edu does allow us to use more language tools than Google does, for the moment. Any list of the most popular words is going to start getting suspect as it moves down the list.