TIME magazine is a treasure trove of information on American politics and culture since 1923. You want to write a paper on it, TIME probably has an article about it. While clicking on article after article on the magazine’s online archive will produce many historical facts and observations, TIME has one source that many people overlook: the actual words used in the articles.
This is where the TIME Magazine Corpus of American English comes in. Created in 2007 by Mark Davies, a linguistics professor at Brigham Young University, users can search the magazine for words from 1923 to 2006.
But first, what exactly is a corpus? Merriam-Webster defines it as “a collection of recorded utterances” that can be used to analyze a language. Instead of using audio to analyze American English in TIME, users look at the frequency of certain words or their context.
This is where the research starts. (If you were confused like I was as to what “KWIC” stands for or what a “POS List” is, the little question marks make a world of difference.) First, chose a display. “List” looks like a spreadsheet, with a column for each decade and a row for the number of times the word was mentioned. If you click on the number, you can actually see the word in context in the screen below and the date it was used.
“Chart” shows the same data in bar graph form, which is more useful for actually visualizing the information to see when the word was used the most. Again, if you click on one of the bars for a certain year, you can see the context each time it was used in the bottom window. This is actually called “Keyword in Context,” or KWIC, which is the third display option. All of the different colors and outline or no outline indicates parts of speech.
The last display option, “Compare,” allows you to compare nearby words for two different words.
Now for the search string. If you want to get really advanced, the question mark next to “Word(s)” shows all the symbols needed to make a simple search more exact. For example, searching [=bright] would give you all the synonyms for “bright.”
The collocates box allows you to search in the surrounding context of the word, and you determine how “far away” the search should take place. The “POS List” allows you to chose the part of speech (POS) you’re searching for.
In “Sections,” you are able to chose the range of dates you want the corpus to search, by individual years or by decades. Finally, in “Sorting and Limits,” you can chose how to sort out the information (frequency, relevancy and alphabetically) and the minimum number of times a word must show up to appear in the results. Choosing “mutual info” for “minimum” will take out common words like “a” and “the.”
It’s a lot to sort through, (and I’m sure there’s something I missed) but all the different options allow you to refine your search, which I’m sure is great if you’re a linguist. If you played around with the corpus enough, you might have noticed that after 10 to 15 searches, you have to register for an account. It’s free and it saves all your past searches, which is really useful.
It’s obvious how linguists can use this site, but what about historians? The words linguists use to track the change of languages can be used by historians to track the change of American society. Take the word “web,” for example. It’s interesting to see where the word changes from being an intricate pattern to the World Wide Web.
Looking at these changes can help us ask new questions. I searched “groovy,” (my dad’s suggestion) and found that the word made a come-back in the 2000s from the 1960s. Why was “groovy” reappearing in TIME magazine? What does that say about our culture in the 2000s compared to the 1960s? What other similarities might there be between the two decades?
What other ways can historians utilize the corpus? Is it a more effective tool when you know what you’re looking for, like my example of “web?” Can you think of any search tools that would make the site more useful for historians? Did you discover anything interesting when playing around with the site?
I have recently grown very fond of the Time Corpus. For the historian it’s an excellent resource to see who was talking about what, when. Although Google N-grams does a similar thing in that you can track literary trends, N-grams just gives you the statistics where as with this you can actually link to the articles themselves.
For that reason this would be extremely beneficial to the researcher.
Also, I’ve found that the “collocate” feature is particular important for the historian who is looking for discussions of certain specific issues. With it you can search words that are only used in relation to other words. For example, if I wanted to search “cake” but only articles that mention people “eating” cake and not “making” it, you can use the collocate feature to make that distinction.
Allen,
Your example of using collocates to find specific issues in TIME made me realize something – if a historian is using the TIME Corpus to essentially search the magazine database for relevant articles, there is no access to the whole article. If you go on the TIME website, it says that full access to the archive of magazines is only available to magazine subscribers.
If historians are using the corpus to look for discussions on specific issues, like you said, and they don’t have access to the TIME archive, they will have to do with the little line the corpus provides them with the collocates. Is this restriction preventing historians from using the corpus? Probably. But that gets us into open access stuff.
I agree with Angela and Allen that the Time Corpus is an interesting tool that offers neat possibilities for scholarly research. As I explored the corpus, I couldn’t help but wonder to what extent are historians actually utilizing this tool? Is it something that a signficiant number of historians are actively using in their research, or does it remain outside of the standard research methods that historians employ? Although it may be difficult to offer quantifiable answers to these questions, my guess is that there are not a great number of historians using this tool on a regular basis. If my guess is correct, what are some possible ways of encouraging more historians to use the corpus? As a class, do we think that this tool should be used more often (or, at least, as often as is practical) in historical research? Or does this tool only offer limited benefits for historians?
As Allen pointed out, the TIME Magazine Corpus is an excellent way to observe “who was talking about what, when” (I can’t think of a better way to say it!). However, I too found that not being able to access the full articles hindered my understanding of exactly what some words meant in different periods. I cannot think of a perfect solution to this problem but perhaps limited access to educators and students would alleviate the problem to some degree. Can anyone think of a better solution?
Historyfan29, your point about the actual use of this tool by historians was something I had not thought of previously. When I think about it, I would have to agree that I do not believe many historians are utilizing the Corpus. I’m wondering, do you think my previous suggestion to offer limited access to educators and students might encourage more scholars to use it for researching purposes?
It is clear that TIME Corpus gives historians an opportunity to analyze words from decade to decade. For example, one can figure out by looking at TIME CORPUS when the media and the country transitioned from using the word atomic to nuclear by looking up both words and seeing the change over time. While this tool may be more helpful to linguistic researchers and professors or sociologists, who are interested in evaluating social changes, I believe that TIME CORPUS offers historians a creative way to gain perspective on historical language. When a historians can see the time period that a word changes that has historical significance, it then enbables them to add substance to their analysis of events over time.