Choosing the Best Words with Google Ngram Viewer

Magnetic word tiles

C.K. MacLeod

Recently, I edited course content for the Teaching LGTBQ Students course on coursesforteachers.ca. The writer used the terms two-spirit and two-spirited interchangeably. I am not the subject-matter expert in this collaboration, so I didn’t feel qualified to decide on the correct term to use. But as an editor, it’s my job to suggest to the writer if two terms mean the same thing. It’s best to pick one and use it consistently throughout.

When you’re deciding on which word to use, Google Ngram Viewer is the tool I turn to. It is a database of 450 million words, gleaned from university library print books that were scanned for the Google Books project (I even found a scan of my Masters thesis on an obscure topic #shiver). Search results from this database of words can suggest

  • which words are more common than others
  • when words may have first emerged in the lexicon

According to the graph generated by Google Ngram, “two-spirit” is the more common term.

Because Google Ngram can only search words from books published up to 2008 (was that when the database was last updated?), I decided to consult a large group of editors for expert advice on current usage. Two-spirit it was.

Armed with this knowledge, I approached the writer with my recommendation.

Applications

In the example above, I used Google Ngram to determine which term is more common. Ngram is also helpful for determining if a word was part of the lexicon at a certain point in history — which is good to know if you’re writing historical fiction. See Carla Douglas’s article Use Google Ngram’s Viewer to Craft Authentic Fiction for more details.

Limitations

No tool is ironclad, and it’s always important to understand a tool’s limits. Sarah Zhang’s article The Pitfalls of Google Ngram to Study Language will help you use Ngram’s powers wisely.

How to Use Ngram

Ngram isn’t difficult to use. Marzia Karch’s article Google Books Ngram Viewer explains the ins and outs of basic and advanced searches, so you can begin using Ngram right away.

Image by Terry Johnston