First up, I never said I wasn't a hypocrite. But Google Ngrams have shown me how much of a hypocrite I can be at times.
I found Ngrams from Dan Pink's marvellous blog. In short, they are a lovely and distracting tool which enable you to search a significant proportion of all the publications in English and some other languages since 1800 and produce scientific looking charts. You type in a word or a couple of words you wish to compare and NGram delivers you a graph mapping their incidence as a percentage of all the words published over time.
One of my first experiments was to compare the incidence of the words 'Training vs Learning' since1800. Looking at this graph it is almost impossible not to to speculate on the non existence of the word 'training' before the industrial revolution, its steady rise with the advent of Taylorian workflow management at the beginning of the 20th century and then its comparative decline with the advent of the knowledge economy approaching the turn of the 21st century.
But there is next to no real science supporting these sweeping assertions. The trouble is I start off with an idea I think will throw something up and then match my story to what I think I see in the data.
Let's try another one, the incidence of "faith, hope and charity" (or the three theological virtues) could be argued to clearly show the constant decline of religiosity in the English speaking world over the last two centuries barring a slight rally after each world war. Again this sounds plausible and might be readily accepted.
But I have done no real work to interrogate this claim or to test anything substantive. The ease with which anyone can now create superficially compelling data groupings to support their claims can only be a bad thing for self directed learners. It is now even easier to create passable nonsense. And it is not as if we weren't gullible enough already...
But I still really like them...
So Google Ngrams, a marvellous but dangerous toy.