#google ngram

I was back-browsing xkcd, as you do, and 1043 is this:

And given the...subsequent developments about Tumblr, I decided to make the equivalent graph today. I eventually found Google Trends, which I don't think is tracking exactly the same thing, but seems pretty compatible:

(And it's surprising to me, how long and how late Tumblr had more google searches than blog. I guess people stopped searching "blog" because everything was a blog?)

But before that I reached for ngram viewer, and got something interesting.

In this case, the interesting thing isn't the comparison; everything except blog is a visual flatline.

The interesting thing is, this search defaults to start from 1800. Why are there any hits at all? What's going on with that bump in the 1920s? (And a smaller one in the 1860s.)

Let's trim the ending of the series. We get this:

(Note this means that "blog" was more popular in the corpus in 1920 than in 2004, which seems rather improbable.)

Ngram viewer is drawing from the Google Books corpus, which is, you know, directly searchable. So I looked for hits for "blog" between 1900 and 1940. And the first valid-looking hit is from page 6 of that classic work of literature, Over-the-counter Brokers and Dealers Registered with the Securities and Exchange Commission as of January 31, 1936:

Ah. Problem solved?

Almost! But now I wanna know what those other two, earlier bumps, are. In 1822 there's an arithmetic book that has stuff like this:

Let me introduce to you. Google ngram.

So i searched for aromantic (among many other terms not shown), and guess what?

[ID: a graph showing an ngram search for three words: aromantic, aplatonic, agender. The line for aromantic fluctuates in the low during 19th century, then spikes upwards after year 2000.]

Are you telling me that aromantic as a term actually appear in the 1800s? Now I'm sus. But I'm also curious. And Google conveniently provided me links to check.

I shouldn't have been so surprised. Do you remember how aromantic is always corrected as aromatic in spell check, and why people assume aromantic is aromatic?

Behold, fellow people:

[ID: a search result showing where the term 'aromantic' has been used in the 1800s. Excerpts include: plant pot - herbs and other aromantic; the flames soon consumed the sacrifice, and on its extinction, which was ludden, an aromantic odour]

Like what the hell. Aromantic odour. Herbs and other aromantic. Aromantic plants.

I have no idea whether aromantic is a typo or archaic spelling for aromantic, especially after reading this line "the flames foon confuned the sacrifice".

The birth of Programming

I love collecting quirks of English where variation hides in plain sight. We know many of the ways that different varieties of English differ, but sometimes there are differences that no one expects, and lead to intense discussion (for example, where your instep is, or which part of your face you use to frown).

Another example of this genre came up when Steph Campisi emailed me while working on a copywriting project for a client in the USA. As an Australian English speaker, Steph had always used hit and miss, while her American colleagues were using hit or miss.

Google n-gram confirmed variation in UK and US English, with and being the preferred form in the UK these days, and or found in US English. This didn’t appear to always be the case - hit or miss was also more common in UK texts in Google’s corpus until a couple of decades ago.

[you can use the interactional version of the ngrram viewer here]

The OED has hit or miss going back to at least the early 1600s, while the earliest hit and miss is 1897, which also fits with the n-gram viewer.

Democracy

Interest in Democracy peaked during both World Wars.