Still the long s was very much alive in 1920s and 1930s in the countries using blackletter.

I used it yesterday in a small title

:-)

Great link, Nick!

I should perhaps have elaborated a bit in my posting... some of what I read about the disappearance of the long s in other parts of Europe noted that Didot's typeface, and other modern typefaces, were designed without a long s.

Thus, when printers used a modern typeface, they were constrained to eschew the long s.

Present-day readers, of course, find the long s very annoying because it's confusingly similar to the letter f. But, naturally enough, we hesitate to conclude that people in the 18th Century felt the same way; instead, if we prefer what we're used to, no doubt they preferred what they were used to.

However, the fact that short s looks more like capital S and is more distinct from f (and can't be confused with any other lowercase letter instead) is an objective fact. Thus, I think it is not unreasonable or provincial to suspect that, even in the 18th Century when the long s was the rule, that a significant number of schoolchildren learning to read and write hated the long s and wished it would go away.

And that many of them might even have carried that attitude into adulthood. (My mental picture has this being more likely to be the case in the lower socio-economic classes.)

And so, suddenly, one has an explanation of why Scotch Roman conquered the typographic world (except for conservative Germany), sweeping beautiful Caslon and even beautiful Baskerville into the dust-bin. (Of course, it helped that the early Scotch Roman faces weren't too bad looking themselves.)

It wasn't that modern roman faces were so much better looking as typefaces that they made Baskerville and Garamond look ridiculous and dated. (Caslon, well...)

No. It was that whenever printers printed something in Caslon or even Garamond or Baskerville, they insisted (those were conservative and traditional times, after all) on putting that dang long s in all the time! So it was not a preference for modern letter forms so much as a distaste for the long s, I suspect, that drove the rapid adoption of Scotch Roman.

And then, fifty to a hundred years later, when the Scotch Romans became anemic and condensed and hideous in form, Alexander Phemister designed his typeface, and Caslon got revived, and gradually the world emerged from the typographic wilderness and we got the Monotype revivals and Times Roman, and are now living happily ever after.

Finally, a rational explanation for how the world got from the beautiful Romans of Jenson and Aldus to the horrid typography of the later 19th Century. Each step was rational at the time it was taken, and the main impetus was not any "mischievous" effect of Baskerville somehow inspiring Didot, it was Didot performing the invaluable service of getting rid of the long s along with the mischief he, rather than Baskerville, happened to be doing to his letter forms.

So you don't think I've made a sensational discovery that will revolutionize how we understand the history of typography?

But you are quite right that improvements in printing technique stimulated innovation in type design. After all, this is very well known in the case of Baskerville: he designed his typeface to look its best under the conditions of the very high quality presswork he performed.

And now that raises another question. Garamond and Bembo are in use today, and no one complains that they are unsuitable to offset printing or photocomposition. What was stopping the typefoundries of the day from making any subtle alterations to existing oldstyle types, therefore, that may have been needed to make them fit in to the new world of smoother paper, and "kiss" instead of "sock" impressions?

One thing about the modern letterforms was that with their severe straight lines, their shapes were simpler than those of the oldstyle letter. And this mattered because...

Linn Boyd Benton hadn't invented the pantograph yet!

Smoother paper fed a demand for mechanical perfection in letterforms - and mechanical perfection was easier to achieve with the modern style of letter because the punches were simpler to cut.

Yes, there were many complicated feedbacks going on. But I still suspect that the long s played a significant role in helping things along - and, because that was seemingly unrelated, it is something that historians would tend to miss... unless the relevant information is brought together, the way it was in the discussion in this thread.

I heard an interview with one of the Google folks who worked on the nGrams project, on (USA) National Public Radio, and he couldn't stress enough that English-language results before 1800 were unreliable, due to low sample size (not a lot of books) and data pollution from mis-dated books (later books that were mis-dated, often because they were ABOUT the earlier time period and someone screwed up). I'm not saying that this affects the cool long-s graph. When playing around with nGrams, you'll often see wacky values in the earlier centuries.