Saturday, September 24, 2011

Ngram view of Meditation

This morning I reviewed 5 million books (5 billion words) looking for references to meditation and forms of prayer. Oh by the way, I had a little help from Google Labs.

It all started with a TED talk that I watched Friday night by Erez Lieberman Aiden and Jean-Baptiste Michel. In this talk they demonstrated the free Ngram Viewer from Google Labs / Google Books. What a cool tool. When you enter phrases into the Ngram Viewer, it displays a graph showing how those phrases have occurred in a corpus of books (e.g., "English") over the selected years.

What word or phrase should I look up first? Since I have had an interest in meditation, I decided to look up the occurrence of the word "meditation" in books from 1700 to 2000.

This is fascinating because there is a distinct peak around 1820, a fairly deep valley between 1920 and 1940, and full recovery by 2000. It is also interesting to note two smaller shoulders around 1930 and 1980. It is not entirely clear (to me) what the Y-axis means (especially when we get to phrases later), but I interpret this to mean that at both peaks (1820 and 2000) the word "meditation" occurs about 10 times per million words examined (0.001 %). The valley is at 4 per million.

So I wanted to see how this compares with the word "prayer."

The word "prayer" has a distinct peak around 1840 that is eight times higher than "meditation" (83 per million), after which this is steady decline to 21 per million around 1975. Notice that there is no recovery after 1940 and only a slight recovery after 1980 to 26 per million. The curve for "meditation" is the same as in the first graph above, but on a different scale.

Next I wanted to compare "meditation" to the words "mysticism," "mystic," and "mystical." Note the timeline starts at 1800. This is a busy graph, so we should look at one line at a time. The "meditation" graph is in yellow here. I find it highly curious that the graph for "mystic" (red) is nearly upside down from "meditation" with a peak at 9.5 per million around 1930 and a steady decline since then, with no recovery. "Mysticism" is lower and parallels "mystic" but with less of a decline after 1930. The word "mystical" (green) is also similar except that it recovers after 1940 and remains steady since 1960.

Since the words are related, it should not be surprising to see a similar shoulder around 1930, but that does makes me curious about what was going on then. It is also interesting that of these three words, only "mystical" recovers after 1940 while the other two have continued to decline. Strange too (or not) that all four of these words are within the range of 3-12 per million. For comparison purposes, the word "centering" is found infrequently until the early 1900s, but since 1920 it has been relatively stable between 1.6 and 2.4 per million, with a slow decline since 1940.

Next I was interesting in comparing and "transcendental meditation" with "contemplative prayer." Notice that we begin with 1900 in this graph and also note that the Y-axis is greatly expanded (showing a much smaller range). The phrase "transcendental meditation" (in blue) does not occur until about 1960, after which it rapidly rises to 0.016 per million or 16 per billion. There is a distinct peak at this point and then it declines to half that or 8 per billion and levels off since 1990. By contrast, the phrase "contemplative prayer" has had a nearly linear growth since 1910 until it is over 6 per billion by 2000. For comparison, the term "centering prayer" was essentially unheard of until 1970 since which time it has had a linear increase to about 1 per billion. I then added the single word "zen" to the search, and was somewhat surprised to find that its occurrence does not dwarf that of these phrases, although it has enjoyed an increase since 1950, the low point of this century.

So I then compared four kinds of prayer: "contemplative prayer," "intercessory prayer," "meditative prayer," and "centering prayer." This graph starts at 1950. Maybe I should not have been, but I was quite surprised at how closely the lines for "contemplative" and "intercessory" prayer line up. It is almost as those they almost always occur together (even in the same sentence). The nadir is not distinct, but occurs around 1970 (2-3 per billion), and both have steadily increased to nearly 7 per billion.

What makes the above observation all the more remarkable is that before 1950 these two phrases did not parallel at all. "Intercessory prayer" peaked at 14 per billion around 1850 at a time when "contemplative prayer" was quite rare.

One might wonder if the incidence of the phrases has anything to do with the incidence of the adjectives themselves. Apparently not. (Range 0-3 per million.)

No comments:

Post a Comment