2 Include the old charts in this answer for easier comparison between 2009 and 2012 corpora.
source | link

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 corpora, but hardly any in the 2012 corpusChart 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 from the question, but with the new 2012 corpus. This is much smoother and no longer has the large dip:

a much smoother chart

 

And another chart withHere's the same words aschart from the other answer, againwith 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

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 corpora, but hardly any in the 2012 corpus

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

a much smoother chart

And another chart with the same words as the other answer, again with the new 2012 corpus. This shows hardly any muft or fhall type words:

hardly any muft type words

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

1
source | link

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 corpora, but hardly any in the 2012 corpus

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

a much smoother chart

And another chart with the same words as the other answer, again with the new 2012 corpus. This shows hardly any muft or fhall type words:

hardly any muft type words