52

In the google "Ngram" search, which allows you to search the usage of words in recent history, I typed in the widely-used internet word "LOL". To my surprise, this came up: enter image description here

As you can see, there was a large usage of the word "lol" around 1810, even larger than the usage today. In addition, I set the time frame back all the way to 1500 AD, and this absolute unit came up:enter image description here My question is, Does anybody know exactly why the usage of LOL was drastically boosted in the 1600s to the 1800s? Thanks!

Semaphore
  • 97,526
  • 21
  • 393
  • 402
Curious Fish
  • 638
  • 1
  • 5
  • 8
  • 28
    Google tend to have a number of OCR errors from very old works, leading to nonsensical words (for the timeframe) like "lol". One example on Google Book is "by the Houses of Parlia-ment" becoming "by the Houses of Parlia- lol- so . . rtiving an Acment" due to the OCR tripping over itself in the margins. – Semaphore Dec 20 '18 at 04:42
  • 18
    This reminds me of reading the Biblical book of 1 Kings in the Aleppo Codex (10th century AD). Where Elijah is taunting rival prophets with the suggestion that their god Ba'al is on a trip somewhere or had fallen asleep or is "doing his business", one of the scribal notes in the margins is the Hebrew letters לול (lamed, waw, lamed)... which through pure coincidence spells LOL. – Luke Sawczak Dec 20 '18 at 05:36
  • 6
    Maybe the Dutch word 'lol' plays a part here. Fittingly, it means 'fun'. :-) – Spiritman Dec 20 '18 at 11:47
  • 3
    Another thing - looking at some of those graphs, esp. the green one (harder to tell for the red since it looks overall quantized, suggesting very small sample size) from around 1800 to 1830, there is an upnotch and while the fine features of the graph suggest plentiful sources, those sudden, vitrually discontinuous jumps almost surely can't be due to a real historic phenomenon. OCR errors really seem like the major source imo. – The_Sympathizer Dec 20 '18 at 15:12
  • 1
    anyway theze graphs give mee the shzzheeghiees... – The_Sympathizer Dec 20 '18 at 15:15
  • 9
    Even a good OCR may have difficulty distinguishing "lol" from the numeral "101". – dan04 Dec 20 '18 at 17:57
  • 2
    The Answer here explains the meaning of the percentages displayed. From the linked "total counts" file, during the big spike there were only a handful of books per year(the last number in each group). 1650,192820,1161,7 1651,540758,2221,3 1652,168692,1023,3 1653,379618,2677,7 1654,36496,256,2 1655,280899,1789,5 1656,688699,3142,4 1657,310453,2551,5 1658,834659,4509,9 1659,543657,2331,3 1660,130457,1085,5 1661,128825,931,3 1662,239762,1471,3 1663,208750,2021,5 1664,290743,2670,6 1665,269608,2689,11 1666,81564,843,3 – Keith McClary Dec 20 '18 at 22:52
  • 1
    P.S. An emailed question prompted me to investigate my Aleppo Codex anecdote above. The question and answer have been posted here on Biblical Hermeneutics SE. – Luke Sawczak Dec 21 '18 at 01:09
  • 1
    If the top of the graph is a square, it's BS. – Mazura Dec 21 '18 at 20:30
  • @LukeSawczak reminds of Bible misprints, such as the Adulterous Bible etc https://en.m.wikipedia.org/wiki/Bible_errata – Gnudiff Dec 22 '18 at 21:59
  • 1
    "Lolling about" is only slightly archaic, even now, I think. "To loll"... – paul garrett Dec 22 '18 at 22:47
  • @Gnudiff The Bible I has as a kid said that Jesus was born in a manager. It led to some interesting debates during Nativity season. – A C Dec 23 '18 at 22:12

4 Answers4

112

In Ngram viewer, along the bottom you can find the actual books that contain the words. By following those links I came across The Changeling By Thomas Middleton, William Rowley whose text looks like this:

The Changeling

So there you have it. "Lol" is a contraction of "Lollio", which looks like a character's name in this book.

In general, I would be extremely wary about Ngram results for

  • Huge timeframes
  • Where language usage, and the language itself, has changed drastically over time
  • Results far into the past, where the corpus size is much smaller and easily affected by a single text
  • Looking for words that didn't exist during a time period
congusbongus
  • 14,497
  • 2
  • 54
  • 114
  • 13
    Plus most of the other results appear to be OCR errors. – Semaphore Dec 20 '18 at 04:38
  • 11
    Is the text corpus for this period so small that a single book would produce such a spike? If so, then it is quite useless. – henning Dec 20 '18 at 09:26
  • 33
    I must say I quite like the idea of a seventeenth century book containing an exchange like - "I muſt truſt thee with a ſecret, But thou muſt keep it." - "LOL" – Pavel Dec 20 '18 at 09:45
  • 36
    @Pavel "I have a wife" "lol, it's too late to keep her secret" – JollyJoker Dec 20 '18 at 11:11
  • 16
    One more bullet for you: o Where the % numbers are really really small (so you're basically looking at a noise graph). – T.E.D. Dec 20 '18 at 17:16
  • @T.E.D. maybe so, but it accurately represents the growth of 'lol' in the 2000s, so it's not necessarily the case here. – TankorSmash Dec 20 '18 at 20:04
  • 2
    @TankorSmash The number of books can change over time though? – Feathercrown Dec 20 '18 at 23:32
  • @henning As T.E.D. points out it's only a jump from something like one hundred thousandth of one percent to 3.5 hundred thousandths. i.e. from nothing to effectively nothing. – Darren Dec 23 '18 at 20:54
  • 1
    Ngrams also display percentage on the Y-Axis. As the sample of books goes down (as it does when we go further into the past), errors like this become amplified as they constitute an increasing fraction of the diminishing whole. – J... Dec 24 '18 at 12:22
38

Congusbongus's answer is very good, and points out that you can find the actual books that contribute to the spike: link. I think "higher prevalence of OCR errors, and very low sample population" is a solid hypothesis.

I noticed one more phenomenon of pre-1800 English text that could contribute to this particular OCR error. See, "lol" has two "l"s, and pre-1800 English text has a statistically very high number of "l"s when run through OCR... because of the long s. I see at least two cases of Google reading "lol" in the middle of a word that actually contains "loſ" with an S:

Here, in the middle of a pretty good OCR, Google reads "loſt his horſe" as "lol't his horfe."

Here, grasping at straws, Google reads "cloſe" as "C'lOl'i.'"

But Google also reads "lol" in a couple of places for "fol." (meaning "folio") and for "for." Maybe there's something about these old typefaces that makes their "f"s as well as their "ſ"s look like "l"s.

Graham's theory of "lol" being commonly found in "Fol lol" and "Tol lol lay," while not directly relevant to your 1500–1700 search, is certainly borne out by the next century's worth of ngram results!

https://i.stack.imgur.com/HZ4lI.png

Quuxplusone
  • 2,285
  • 1
  • 15
  • 25
  • I thought to test my hypothesis by searching for unlikely words without Ls. The prevalence of kek was higher than I expected, thanks to the legit use of the word in Dryden's Mr. Limberham (1) but mainly thanks to one particular filigree design (2, 3)! And even here we find a mis-OCRed long S: "kek" for "ſeek". – Quuxplusone Dec 25 '18 at 21:16
21

In addition to the previous answer...

"Lol" is a nonsense syllable frequently used in mouth music. Other similar syllables are "fa", "la", "fiddle", "diddle", and so on - think of the song Deck the halls, for example, or Whack-fol-de-diddle from the Dubliners. This has always been a staple of English folk music. More recently, scat singing uses the same principle in jazz. The concept is the same though - the nonsense syllables are chosen for their sustained and/or plosive qualities to fit the meter of the song.

In the 16th century, madrigals spawned a new genre called balletto which featured this as a standard part of the form, and this coincided with the widespread availability of printing. It is not inconceivable that transcriptions of broadsheet songs would skew the statistics for these mouth music "words".

Graham
  • 2,439
  • 1
  • 13
  • 17
2

"To my surprise, this came up"

Why, you don't honestly believe it was an acronym for laughs out loud in 1800 do you?

You should give a little thought to the idea that what you've searched there includes real words as well as acronyms with completely different meanings that just happen to be spelled the same.

An old or alternative (or just lazy) spelling of loll seems the most likely cause of those statistics to me.

meaning to "sit, lie, or stand in a lazy, relaxed way."

Words don't just fall in & out of favor with time, they change, old words disappear, new words appear & something spelled the same way in 1500 as a word in 2018 may have absolutely nothing to do with the current word either in it's usage & meaning or the etymology of its origins.

Your results from 1500 are pretty pointless in this instance & certainly don't point to the use of lol as an acronym for laughs out loud in that century that you seem to be wondering about.

Pelinore
  • 137
  • 5