Tell me more ×
English Language & Usage Stack Exchange is a question and answer site for linguists, etymologists, and serious English language enthusiasts. It's 100% free, no registration required.
up vote 94 down vote favorite
13

According to the following Google Ngram, in the U.K. the modals should, shall, and must were virtually missing from English writing during the 18th Century (I've added will for a comparison modal which was unaffected).

I have never seen this mentioned anywhere, and I couldn't find it in a brief web search. What happened? Was this a real phenomenon, or could this somehow be an artifact of Google Ngrams? Were these modals absent from speech, or just formal writing? More interestingly, how did they get reintroduced? A brief search shows that Shakespeare definitely used these words quite frequently.

enter image description here

share|improve this question
6  
What a fascinating question! – hippietrail Aug 28 '11 at 20:22
3  
@PeterShor: NGram is a fine tool, but what about a more mechanistic interpretation or the data... It seems to me that there is continuous decline the use of all these words after 1825 which is an absurd. What has replaced their use? What has come up in lieu of "will" or "shall"? Do people use the future tense less and less? (Maybe it suggests a manifestation of massive pessimism... ) – Pantelis Sopasakis Sep 13 '11 at 21:00
See these two brilliant articles from 2006 that exploit this "feature" of Google Book Search and Google Ngrams: The Rules for Long S and The Long and the Short of the Letter S. – ShreevatsaR Sep 23 '11 at 4:56
@PantelisSopasakis All modals have been declining, perhaps more markedly in print because of increasing colloquialization. Will/shall have been partially supplanted by be going to and by contractions with 'll, must by have to, should/ought to by be supposed to. – StoneyB Feb 12 at 0:02

2 Answers

up vote 16 down vote accepted

This was a problem with Google's optical character recognition (OCR) mistaking the long s (ſ) as an f.

However, Google has since improved their OCR:

When we generated the original Ngram Viewer corpora in 2009, our OCR wasn't as good as it is today. This was especially obvious in pre-19th century English, where the elongated medial-s (ſ) was often interpreted as an f, so best was often read as beft. Here's evidence of the improvements we've made since then, using the corpus operator to compare the 2009 and 2012 versions:

Chart showing many pre-1820 beft results in the 2009 corpus, but hardly any in the 2012 corpus

Here's the original chart from the question, with the 2009 corpus:

Strange dips

Here's a chart with the same words but with the new 2012 corpus. This is much smoother and no longer has the large dip:

a much smoother chart

Here's the chart from the other answer, with the 2009 corpus:

Strange dips

And here it is with the new 2012 corpus. This shows hardly any muft or fhall type words:

hardly any muft type words

share|improve this answer
1  
Thanks for the update! – prash Nov 24 '12 at 15:44
I’m glad to see that they now recognize ſ U+017F LATIN SMALL LETTER LONG S as an alias of the regular round s, although that does mean you can no longer search for it separately. But I’m quite displeased with their lame scan of Sir Frederick Madden’s translation of Laȝamon’s Brut, as it (next to) never correctly processes any of the thorns, yoghs, or eths. – tchrist Dec 3 '12 at 13:37
See the conversation in chat beginning here for how they only manage the long-s, and nothing else. – tchrist Dec 3 '12 at 14:31
up vote 73 down vote

Interesting! I suspected the reason was that the character s used to be written somewhat like the present-day f. (It would look like s only when it was the last character of a word.) I looked at some of the OCRed documents that Google Ngrams links to. That only strengthened my suspicion.

Look at this plot (also below). I have plotted for "should,shall,must,will,fhould,fhall,muft".

We had some very important literary works between the late 1500s and the early 1600s. I guess these have been re-published with the old-fashioned s replaced, and therefore, OCR worked better.

added fhall,fhould,muft, and extended the range

share|improve this answer
2  
It looks like this is right! ngrams.googlelabs.com/… shows that the fall of should is accompanied by an equal rise in fhould, etc. – Jeremy Aug 28 '11 at 15:03
17  
@Prof. Shor: One important (to me) consequence is that it shows why we can't rely on (just) Google NGrams to settle questions about early Modern English. – prash Aug 28 '11 at 17:16
7  
Great answer to a great question. The version of "s" that looks like "f" is called "long s" and Google's OCR definitely doesn't detect them at all. I tried with ſhould,ſhall,muſt and got zero. – hippietrail Aug 28 '11 at 20:30
1  
@prash can you include the plot in your answer? that way both images can be compared from this single Q&A. – DuckMaestro Aug 28 '11 at 20:33
1  
@DuckMaestro: I now figured out how it's done! – prash Aug 28 '11 at 20:41
show 8 more comments

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.