Popularity of the canoe versus other outdoor activities over time

I have always understood that canoes were hugely popular during the late 1800s and early 1900s when compared to similar outdoor sports but I've never found a good way to measure this. The Google books Ngram Viewer as described at http://books.google.com/ngrams/info provides an interesting approach to this question. This gives a way to count the occurrences of words in books that have been scanned by Google as compared to the time when those books were published. The results at http://books.google.com/ngrams/grap...start=1800&year_end=2008&corpus=0&smoothing=4 show how frequently canoe, bicycle, ski, golf, and kayak occurred in books between 1800 and 2008. There is an obvious peak for canoe between about 1850 and 1940 and usage has also been rising sharply again since 2000. The usage of words related to other common outdoor activities seems to match their popularity. There are probably many cases where the word "canoe" is used that don't have anything to do with current outdoor sporting activities but this does provide an interesting perspective.

Benson

Interesting and encouraging to see the recent upturn.

On a side note, I'll be spending tomorrow afternoon in a roundtable discussion RE: promoting canoeing amid a high-kayaking community. Not really sure where it'll end up, but it'll be interesting...

Paul, bring your boxing gloves!

No way, man... the gloves are off!

Canoeing is more likely to include another person. Also, you don't have to get your butt wet. I think canoeing is more civilized in that you sit up with good posture instead of reclining in your craft. Also quieter and more elegant.

If I change the corpus to American English to restrict the database to books published in the US, the curve changes shape a bit, dropping off faster around 1920 and not recovering in later years. This pretty much agrees with my own pre-conceived ideas based on what research I've done, that the golden age of canoeing started around 1860 but started to wane in the late '20's and into the 1930's.

The puzzle for me is the early peak. If MacGregor really kicked off the canoe culture in the 1860's, that would explain the rise in the curve about then. The puzzle is the peak around 1850. Perhaps we are dealing with a fascination with Native American canoes? The Last of the Mohicans was first published in 1826 and its sequel 14 years later. If you restrict the books to English Fiction, the first peak is there and there is a gradual rise up into 1920 or so.

Unfortunately, not knowing the details of the database they are using it's hard to put too much meaning into the details. (or perhaps it's too easy to read too much!) The curve should probably be normalized somehow for the number of books in the data by each decade... or maybe it already is.

Lotta fun in any case. A great thing to play with.

Thanks for making us aware of this.

It is interesting to compare the corpus differences since the recent rise after 2000 appears to be almost entirely based on English fiction. This may be due to the British use of canoe for most small boats including the covered ones that are generally called kayaks in North America. They tend to use Canadian canoe for the small open boats that are simply called canoes here.

I agree that the early rise before MacGregor started publishing in the 1860s is unexpected. This is probably part of a general interest in Native American culture as you suggested.

Benson

The other element is the American obsession with the Great Outdoors as promoted by painters such as Bierstadt. The rising conciousness that Wilderness represents a philosophical entity and the recognition of that loss due to sheep farming and lumbering could also contribute to the rise in popularity of canoeing. The blip in art produced "en plein air" in Europe may symbolize a leap in human evolution, why not canoeing?
Shirley Reid

When I entered the term what should be a fairly constant term: "man" there is a similar trend between 1900 to current. So I think that the results need to be normalized to some term that would have a constant frequency of occurrence. I think that the number of words in the dataset declines through the 20th century and picks up in the 2000's.


Paul

I tried a similar exercise with a number of "average" words and got mixed results, but often a dip mid-20th century. I think in broad brush the results are meaningful. You just need to be cautious about interpretation.

I'll blame the early peak on those darn "arts".

If you've gotten hooked on the nGram viewer, then the following YouTube link is required viewing. It's from the Boston TED conference and is a talk by the originators of the project.

http://www.youtube.com/watch?v=SdhJwLJcdsU

ACA new memberships by year

Some more trivia.

I recently came across a PhD thesis for PE: "The History of the American Canoe Association, 1880-1960". There's not a lot of insight in it, but the fellow really did comb the old yearbooks.

One thing that caught my eye was a table of membership numbers issued for each of the years in this time period. As the ACA issued sequential membership numbers, (and still may for all I know), with a little math you can calculate the number of new members each year. I broke down the results into 5 year chunks and the resulting graph is (I hope) attached to this note. It sort of mirrors the results of the n-Gram viewer, but whith a drop off starting a little earlier, around 1915. This also conforms my subjective observations that by the 1920's the golden age of canoeing was on the wane. The dip in the 1940's must be the war years.

Note that this is only NEW members each year. You can't get total membership by year from these numbers.

That's interesting... and it's good to see that your efforts are uncovering some of the information you were asking about, a while back.

ACA has changed its numbering system a couple of times recently... My original number (ca. 2004?) had six digits, then they went to a nine digit format, and now it's 8 digits. The numbers do not resemble each other in any way. So, at least for the last seven years, the numbers are not sequential... they seem to be randomly generated...

fwiw.