12

Google Ngram viewer allows one to compare the frequencies of a set of phrases over time. It even allows you to restrict that comparison to an American corpus, or separately to an English one.

What I am wondering if there is any way to compare the frequency of one word/phrase over time between American and English? Either through Google Ngram or something else. I couldn't really find anything on google (itself or by using it).

1
  • 1
    +1. Mitch, can you edit the title to say 'on the same graph'?
    – smci
    Oct 15, 2013 at 23:22

2 Answers 2

Reset to default
6

Through lack of recent practice I'd forgotten how to embed a Google Ngram chart in a post, and in my search for where I'd read it on Meta - which I finally found here - I came across this old question. As Google Ngrams now have substantially added functionality, I thought it would be worthwhile to post an updated answer.

You can now easily compare frequencies between the British and American English corpora in the one chart. In the "comma-separated phrases" field, insert the : corpus selection operator after the search term, followed by the shorthand for the corpus (e.g. eng_us_2012), add a comma and repeat the search term with section operator and the shorthand for the other corpus (e.g. eng_gb_2012). You can do this with more than one search term, to see which appears more frequently in each corpus.

As an example, compare the expressions take a walk and go for a walk. A basic Google Ngram using the search {take a walk,go for a walk} shows that the former has always occurred more frequently, but the latter has been growing in usage:

"take a walk" vs "go for a walk"

However, if I wanted to look at transatlantic differences in frequency, I'd search for {take a walk:eng_us_2012,go for a walk:eng_us_2012,take a walk:eng_gb_2012,go for a walk:eng_gb_2012}. The resulting chart reveals that take a walk has declined in frequency in the GB corpus over time, to such an extent that go for a walk has been the more common term (albeit narrowly) since about the mid-1930s:

"take a walk" vs "go for a walk" for U.S. corpus vs G.B. corpus

The Google Ngram Viewer info page provides not only the shorthand codes for the different corpora, but also a summary of all the other useful "Advanced Usage" options, including:

  • Wildcard search (shows the top 10 substitutions for the word, e.g. University of *)
  • Inflection search (e.g. "book_INF a hotel" gives results for "book", "booked", "books", and "booking")
  • Part-of-speech Tags (really useful! E.g. differentiate tackle_VERB vs tackle_NOUN, or find all expressions of the form "read (determiner) book" with read _DET_ book)

and a range of even more advanced and complicated search functions!

4
  • 2
    I don't trust Google NGrams for this (nor many of the advanced functions that it has), for the reasons I outlined here. A basic search shows that "colour" occurs only twice as frequently in the GB corpus as "color". This is because what corpus texts are sorted into is based on their place of publication, not who they were written by or some other metric.
    – Laurel Mod
    Jan 10, 2019 at 2:55
  • 1
    @Laurel yes, I understand your distrust (I read your answer when I was searching through Meta). This question has relevant critiques too, though it applies to the "old" Ngram Viewer. Certainly the Ngrams need to be used cautiously, but they can be useful as broad indicators: it just needs an awareness of how a specific search might be flawed (e.g. your colour example).
    – Chappo Hasn't Forgotten Monica
    Jan 10, 2019 at 3:19
  • 1
    @Laurel I also like Sven's comment: "The primary value of Ngram to me isn't that it draws trustworthy lines representing the frequency of use of a word or phrase on a timeline; it's that it helps me search for actual occurrences of a word during a particular period of time."
    – Chappo Hasn't Forgotten Monica
    Jan 10, 2019 at 3:19
  • @Laurel one of the reasons for the odd GB result for "colour" is that a US book might be republished in GB. Have a look at the Introduction (p.7) of Color and Meaning, first published by University of California Press then this edition by a UK publisher. The text is an extraordinary mishmash of US and GB spellings: it's spelled colour in general text, but retains the US "color" when quoting (frequently!) from sources. This one book on its own could skew the GB result considerably!
    – Chappo Hasn't Forgotten Monica
    Jan 10, 2019 at 4:17
3

On Google's Ngram viewer you can set the corpus to be American English or British English, and get a graph for each. You can then compare the y-axis values, being careful to note that Google autoscales it.

For example: American English and British English.

You can also download the datasets of each corpus if you'd like to do your own data processing.

[source]

6
  • 2
    The scaling is the problem. Having both on the same graph is what I'm looking for.
    – Mitch
    Oct 31, 2011 at 23:46
  • 1
    @Mitch I've spent some time since you first asked about this, learning about ngrams. Hugo is correct, I believe. The Ngram Viewer if for fairly simple inquiries according to Google. In order to get what you want (one graph with both series of data, possibly using dual y-axes for scaling if necessary or even only one y-axis), would require running the datasets yourself.
    – Ellie Kesselman
    Dec 15, 2011 at 18:37
  • 1
    The URLs should be updated, I think. They redirect correctly, but Google has since closed GoogleLabs, so googlelabs dot com doesn't exist. The first URL is now books.google.com/ngrams/… and the super relevant one (my opinion) for datasets is now books.google.com/ngrams/datasets
    – Ellie Kesselman
    Dec 15, 2011 at 18:41
  • Thanks for the update...I'll see what I can do with the download.
    – Mitch
    Dec 15, 2011 at 20:10
  • @Feral, thanks, and remember you can edit posts yourself to fix links. Mitch: it's a LOT of data. The page gives an example of line 30,000,000 in a file, and there's 10 files per corpus - and that's just the 1-grams. The larger grams have many more files. Still, it'll keep you busy for a while!
    – Hugo
    Dec 15, 2011 at 22:11
  • I wonder if we should have an ngram tag? @Hugo I can't edit on meta, not sure why.
    – Ellie Kesselman
    Dec 15, 2011 at 22:20

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .