Thirty Dashes

My twitter stream informed me the other day that “-30-” was often used at the end of newspaper stories, to mean “end of the newspaper story”.  It's something I'd seen once somewhere before, but I hadn't really given it a second thought.  But the tweet linked to an article about possible origins of this story, and so I, being the eternal doubting Thomas, had to jump in and take a look.

That blog put forth a few possible reasons, none of which panned out.  I find no sign that thirty came from “XXX” for end of story, or that AP writers had a quota of 30 stories, or that 30 was a typesetting symbol.  I also find no evidence of a news editor named “Thirtee”.  A favored theory seems to link the use of 30 to Civil War press telegraphs.

Then I found this, which I thought would settle the issue.   It was a reference to the use of “30 dashes” in this 1936 book “Newspaper Makeup” by John Edward Allen:

“Many newspaper makeup men make the mistake of putting the same amount of leading above ‘30’ dashes (the dashes at the ends of stories) as below them.  Some makeup men actually put more leading above than below.  As ‘30’ dashes are supposed to end stories, not merely to separate them, and surely not to form parts of following heads, it is advisable to put more leading below than above them.  And if body lines of stories are set on slugs of larger point sizes (if the lines themselves are cast leaded), that leading on the slugs themselves should be allowed for when extra leading is placed above ‘30’ dashes.  More leading should be used above ‘30’ dashes when body lines are cast solid than when cast leaded.  But in either case more leading should be used below the dashes than above.”

“Leading” refers to the use of leads (wide bars of metal, originally lead), to add space between lines of print.  But  more to the point, “dashes” here does not mean the dash or hyphen or minus sign, although these were all called dashes.  A dash in this context was like a rule.  A rule is, of course, an ornament used in a form that would print a line from end to end.  A dash on the other hand would print a line (or some sort of separator) in the middle of the space, and leave the ends blank.  I'll put a dash in here.

That's a fairly ornamental dash.  Newspapers were more likely to use a simple straight line as a separator between stories.  So I thought I had my answer, that the notation simply meant “put a 30 dash here” (which it certainly did) and I assumed that “30” was going to be some sort of typographic measurement.  Dashes were certainly measured in ems, and a 6 em dash was a standard thing.  Maybe this was a 30 em dash?  Nope, newspaper columns were fairly standard at about 13 ems in pica (today we'd call it 12 point) font, so 30 didn't make sense in that context.  But dashes were also assigned numbers based on the thickness of the line, and I found 27s and 33s.  But no 30.

But then I found this from 1919:

From Telegraph and Telephone Age, volume 37, November 1, 1919.

I also found a more detailed article from 1906.  It agreed with 30 as a “collect charges” code, and also  mentioned other theories, some similar to the blog summary I linked to above, including a person named Thurtee, and claims of civil war origins.

But I was still a doubting Thomas.  An 1888 story about “30” and other signals made no mention of the idea of collect charges.  So I looked for the evidence that “30” used to mean “collect charges” at some point in time.  After much googling, I came up empty.  If anything, I found that “pd” and “coll” were common abbreviations used at the end of messages to indicate the charging status.  But I did find that there were standard “signals” used in telegraphy, dating back to at least 1854, where I have found two different (but similar) sets of signals.  These signals were numeric codes used to transmit simple, common messages like 1 for “Wait”; 2 for “What time do you have?” (which was really important back before we had time zones—which made railroad schedules a nightmare, and which is why the railroads invented time zones for us) and 12 for “Do you understand?”.  While all of the signal sets I've seen have similarities, it's also clear that each company used their own set of signals, and some texts specifically call them “private” signals, because there was no standard.  By 1862, 30 was a code commonly used, at least in railroad telegraph systems, to mean “Finis”.  A very similar set of codes was later adopted as the Phillips Code in 1879, to be used by the Press in their telegraph transmissions.

From The Christian Union, volume 37, May 24, 1888

A few modern web sources and recent books (later than about 1960) claim that there was a “92 Code” introduced by Western Union in 1859.  But I have so far found nothing to substantiate this name for a signal system, and “92 Code” is probably a modern label.  Certainly the old texts I've found all call them “signals” or occasionally “abbreviations” or just “numerals”, but never codes.  And while 92 was the largest number used in a few of the systems, other older systems used higher numbers, including 134 which commonly meant “who is writing this?”.

I've also found no evidence that Western Union introduced any set of signals in 1859, however it does seem likely that they had a set of signals in place by then, if only because I've found other older signal sets.  And of course the “magnetic” or “electric” telegraph was preceded by semaphore telegraph systems, which also used signals for common messages.

Despite some modern claims, these codes were only used for conversations between telegraph operators, not for substitutions of text within the telegrams.  Most companies had rules forbidding this to ensure accurate transmission, and in fact numbers themselves were often transmitted as text, or double transmitted, once as a number and once as text.  I found at least one signal which meant “Write the figures into words”.

Another common code was 73, which meant “Best regards”.  Some modern sources suggest that this was originally “My love to you” or “love and kisses”.  Other sources put 88, or 22 as “love and kisses” (and one even claims this was the “first emoticon”... UGH!).  While these usages may have been popular during the ham radio years, they seem unlikely during the early telegraph years, given that they were signals sent between telegraph operators, and not within telegrams.  Some stories trace these claims to two sources that I haven't reviewed: “The National Telegraphic Review and Operator's Guide” from April 1857, and the “December Bulletin from the Navy Department Office of the Chief of Naval Operations” from 1934.  I did find the July 1853 “National Telegraphic Review”, in which there is an annotation on a poem, where telegraph codes were used in this poem.  The annotation translates 73 as “My love to you”, but this may simply be what it best represented in the context of this poem.

A portion of a poem from The National Telegraphic Review and Operator's Companion, July 1853

(That same poem also includes “O. K.” in quotes, as if it might be a telegraph notation.  It was, and was used to mean what it means today.  It seems that it was a popular joke in 1839 to say that “O. K.” was an abbreviation for “all correct”.  Some people point to the telegraph as the reason this joke survived long enough to become standard usage.)

It's clear that numerical codes called “signals” were in wide use very early on, and it's also clear that there was a lot of variation in the codes.  “30” became so common that it by 1888 it was worthy of mention in mainstream sources, and also by that time, its origins were too old to clearly piece together.  In my opinion it was just an arbitrary choice, like any of the other signals in use.  But you never know—history can be a hard thing to pin down.

—73—

—30—

Show The [manipulated] Data

Myths are powerful factors in our lives.  They can blind us from the truth.  One such myth I've discussed often is the idea that monospaced fonts demand more space between sentences than proportional fonts.  Any rational analysis of the geometry will completely contradict this myth, as does the history of what actually happened when typewriters were introduced.

But myths are so powerful that otherwise rational human beings will literally lie to themselves to hang on to the myth.  I came across one such example today on a blog called "Show The Data".  It's a blog about the visual display of information, and they created a visual in service of the monospaced font myth.  It's a striking visual and at a glance forms a compelling argument for that viewpoint, until you begin to notice just how hard they had to work to distort the data, rather than showing it.

What the image seems to show is that in a proportional font you automatically get more space between sentences, while in a monospaced font, your sentence spacing is too similar to the space between letters (let alone words).  And indeed those pink rectangles are more similar in the monospaced font than in the proportional font.

If you like the myth and you like the visualization then at this point you're done and move on.  And you'll never notice that you've just been fed a huge lie in the form of visual distortion.

Look again.  The pink square between the two “i”s is not really between them, it's on top of them.  It completely overlaps the serifs on both letters, and touches the stem of the letters.  This essentially doubles the size that this pink rectangle should actually be.  Now look at the pink rectangle between the period and “Maria”.  It does not touch the period, and it does not touch the “M” (to say nothing of overlapping the serifs).  And now compare that pink rectangle to the other pink rectangle in the second line.  This pink rectangle does overlap the serifs of the “M”, and touches the stem.  It not only touches the period, it completely overlaps it, and covers whitespace to the left of the period, almost touching the “g”.  So what happens if we fix these problems in the image, and consistently mark the spaces?  I adjusted the boxes so they all touched adjacent serifs without overlapping them.  I chose to cover both periods entirely because in my opinion periods don't have enough weight to take up real visual space, so including the period in the spacing, as he did with the proportional font, is the way to go. This is what we get:

That's a striking difference from the first graph.  The argument looks ridiculous now.  It's even worse actually because he totally cheats on the space between the period and the M in the proportional font.  The space between “Maria” and “was” is nine pixels, but he has twelve pixels between the period and the "M".  He probably chose one of those odd fonts that misguidedly adds space to the right side of the period regardless of how and where it is used.  This means that he has exaggerated his idea of sufficient even more than I've shown here.

But you might have noticed something else that's odd.  He compares a letter space in the first line with a word space in the second.  He sort of covers this issue in the blog text, where he talks about the problem you can have with letter spacing and how it can confuse word and sentence parsing.  This actually plays to my advantage here, as it makes the comparison ridiculous, where if I were comparing word and sentence spacing between these two lines they would actually both be very close to double the size in both fonts, as a period and space is twice as big as a space in all monospaced fonts and very close to that in most proportional fonts.

Still, if he was simply trying to make a point about letter spacing issues in monospaced fonts, why did he, an alleged expert in the portrayal of visual information, have to cheat so hard to make that point?

I haven't talked to the writer, but I'm going to guess that he did not intend to commit a fraud upon his audience.  Most likely he set out to demonstrate a point, and when it didn't work he fudged reality to fit the myth, rather than questioning the myth itself.

Because we're human beings.  And that's how we roll.

April Fools' Day Origin Stories

A tale of rape and prostitution, religion and science, fish and birds.

Note: Even though I am publishing this on April Fools Day, and about April Fools Day, I swear it is all true.  Well, not so much true as accurate.  That is, stuff here may be made up, but it wasn't made up by me.  But, because I'm rushing to get this out today, I won't bother with adding any sources.  Or with things like poof-reading.

While some sources say that April Fools Day, sometimes also called All Fools Day, began around 1700, it's pretty clear that it's older than that.  But in the 1700s and 1800s April Fools Day was all about “making an April Fool”, that is turning a person into an April Fool.  One source bragged that someone had “made over a hundred April Fools” (not a direct quote, just from memory).    In other words, he tricked over a hundred people.  While today, the trick can be all sorts of things, at that time the focus seemed to be almost exclusively sending people on foolish errands (and may well be the origin of the “fool's errand” as a common idiom).  (Note: I just made up that connection about “fool's errand” and while it could well be true it does mean I lied up at the top there where I said I wouldn't make anything up.  Sorry.)

It was apparently such a big problem that some sources describe the difficulty in getting people to perform errands on April first, on the grounds that it was probably just an attempt to make them into an April Fool.

One claim I see in a few sources is that April Fool's Day came about because of the change in Europe from the Julian to the Gregorian Calendar.  New Year's Day was apparently frequently celebrated on March 25th or thereabouts and somehow this means that people who were celebrating the New Year on this schedule were out-of-date, and therefore fools.  There's a few problems with this.  For one, there seem to be sources that refer to the April Fool custom before the Gregorian calendar was adopted (for example, a 1539 Flemish source that I read a convincing description of somewhere).  The other problem is that under the Julian Calendar, New Year's Day was actually January first also.  The practice of celebrating New Years Day in March was a local custom that was just as out-of-step with the real calendar before this change as it was after.

Clipped from Theorica Della Compositione Dell' Vniverso Et Delle Cavse Della Nvova Riforma Dell' Anno, 1582
It has something to do with the Gregorian calendar.

Just to confuse this issue, there is another holiday, called “The Feast of Fools”, which traditionally took place on January first.  And possibly before that on November first.  And then there's also a “Feast of the Ass” that existed at some point, perhaps on April 12th, but it doesn't seem to be connected.

The Rape of the Sabine Women, Sebastiano Ricci, ca. 1700
(There are a bajillion paintings of this event.  Seriously you say "rape" and painters are like "I'm there".)

A few different sources (which are not parallel and so it could be case of later sources all quoting one single earlier source) seem to say that this tradition arose in response to a bit of Roman history: the Rape of the Sabines.  (In which “rape” meant that they were abducted to be taken as wives, and some sources make a point of clarifying this different use of the word “rape”, but I'm figuring if they were abducted to be wives there was probably something happening that in modern terms we would still call rape.)  Apparently this is a real, true bit of Roman History (which could of course be a myth), in which some Romans convinced the Sabines that they were going to have a sort of Olympics for the god Neptune, and why don't you come here and compete and bring all your wives.  And they did, and the Romans killed the men and took the women.  And this occurred on April first, and today we celebrate the rape and kidnapping and murder by playing cruel jokes on each other.   For some reason this origin story doesn't seem to be very popular any more.

"Scomber Le Maquereau", Poissons de Mer, Aalbert Flamen, ca. 1660

The next one isn't so much an origin story, as a rambling study of French slang.  Apparently the expression in France for “April Fool” is “poisson d'Avril”, which literally means “April fish”.  It's pretty clear that the April Fool tradition has been in France for quite a while, and is one place to look for a real origin story.  Certainly we find this expression “poisson d'Avril” in sources going pretty far back pretty easily.  But this could be because the expression also means “pimp” (or as they said back then “bawd” or perhaps “pander” which allegedly comes from the Shakespeare character Pandarus, although as with any origin story it could be older than that and Shakespeare was referring to the word in the naming of the character.  Who knows, history is a mess.)  This connection could come about because another french slang word for pimp was mackerel (“maquereau” or “maquereaux”). And the same expression also seems to mean effeminate man (and in that old sense might have been a way of referring to gay men.  Yes this is me making stuff up again.)  This seems to be because mostly women ran brothels, not men.  At any rate, the jump from “mackerel” which was a fish caught in April to “April fish” as alternate terms for pimp is not so great.  None of which clarifies the connection to the tradition of foolish errands at all, although it certainly could be related to the ideas of tricks on April Fools Day, and the modern phrase “turning a trick” which is a term related to prostitution.  (I'm completely making this connection up though.)

From Histoire de France. Le Blog La France pittoresque (Not sure about actual origin.)

Despite that dual usage, it's still pretty clear from context that only some of the old usages are meant to be “pimp”, while others clearly refer to April Fools Day traditions.  So France remains in the running as a possible origin for this tradition.

At least one source says that “poisson” could be a confusion of the word “passion”.  Unfortunately this has nothing to do with the movie Passion Fish (as far as I know), so there's no excuse for me working that in here.  But the suggestion here, which actually does relate to April Fools Day, is that this is a reference to the Passion of Christ, with Jesus cast into the role of the fool, as he is bounced around Jerusalem from one authority to the next on a wild goose chase (or, one might say, a “fool's errand”.)  This of course all took place sometime in the vicinity of April first.  Being an old source, they naturally blames all this on the Jews rather than the Romans.  Therefore, April Fools day was cast as a Jewish celebration of the success in tricking Jesus.  In which case April Fool's Day is basically a leading cause of the Inquisition.  (Note: that last remark was 100% made up by me, and bears no relationship whatsoever to reality.)

In a similar vein, some have blamed April Fool's Day on the Jews by tying it to the fool's errand that Noah (you know, the one with the flood) sent that first dove on, before there was any land to find.  Which conveniently makes a nice segue between religion and birds.

One of many types of cuckoos
From Histoire naturelle des oiseaux d'Afrique, vol. 5, 1799, Le Vaillant, François

Yet another story claims that April Fools Day can be blamed on the cuckoo (or cuckow) bird.  The cuckoo is a brood parasite — a bird which is too lazy to raise its own kids, so it lays them in the nest of some other bird, and lets that bird raise them.  Just to add weight to this theory, in some parts of England an April Fool was called an “April gowk”, where “gowk” is another word for the cuckoo.  The birds typically lay their eggs around April First, and so somehow (don't ask me how, I didn't make this part up), this leads to a holiday in their honor.  A holiday to celebrate child abandonment.  Which goes right along with the theories about rape and prostitution.

So the majority of these theories seem to be oppressive of women, either tying the holiday to rape, prostitution, or a mom abandoning her child.  And some of the other theories seem to be anti-semitic.  Does this mean that we should reject this holiday, as it is clearly about oppressing somebody?

Or does it just mean that people that invent stories to fill in missing history are jerks?

Maybe we're fools for falling for it.

[But seriously, that closing sentence was not some sort of “gotcha” admission of this all being a trick.  This stuff is all real, I swear.  Except for the parts that aren't that I mostly mentioned in there.  Mostly.  I really did look at a whole bunch of sources.  Here's one that's an excellent summary, just to make you feel better: Popular antiquities of Great Britain, 1877, sir Henry Ellis.]

The First Car Advertisement?

This ad appeared in Scientific American, v79 No. 5, July 30, 1898; via Wikipedia

People absolutely love “firsts”.  The first man to walk on the moon.  The first women to swim the English Channel.  But history is rarely so convenient.  So when I see some historic “first” mentioned, I instincitvely look for some sources, because more often than not, it isn't the first.

The ad may have a slighlty legitimate claim, as long as we pile on a few caveats.  It may be the first car ad, selling a specific and real model of car that was in production, and had a set, advertised price, in the United States.  It is certainly the earliest one that has been widely reported.

The Winton Motor Carriage Company continued advertising in Scientific American throughout the rest of the year and (and onward).  They used very similar ads in every subsequent edition (published weekly) that year.   The following week, August 6, the same ad ran with the headline “Better than a Horse or Bicycle”.  Sometimes the ad reused headlines but mostly each week was a new headline on the same image: “The Luxury of Locomotion”, “The Winton Motor Carriage”, “A Delightful Drive”, “Over the Hills and Far Away”, “In Season and out of Season”, “Another Lot of New Ones”, “You Are Invited”, “The Other Fellow”, “Snow, Ice and Slush” [December 10, and don't blame me for the lack of an Oxford comma], “Wonderful Control”, “The Automobile”, and “The Proof of the Pudding” to close out that year's ads.

They were pretty small ads, appearing in the advertising section at the back of each Scientific American.  Here's the ad in context (if you can even spot it):

page 80 Sci Am v79 #5, via hathitrust
The car ad is second from the top in the first column.

The image was tiny, which partly explains why an engraving was used rather than a photograph.  Photographs were fully viable by this time, and if anything the 1890s were a boom period for photographs in print as much as they were automobiles.  This very Winton car image from the ad had already appeared as a photograph in Scientific American just a couple of months earlier:

Scientific American v78 No. 20, May 14 1898, via hathitrust.

The photo was much larger, covering nearly half the page, along with an article titled “The Winton Motor Carriage” that took up half the text in the remaining space.  “A ride in a motor carriage is a comparatively new and delightful sensation” the article begins.  It goes on to mention that Europe is ahead of the U.S. at this point in the car game, with the “motor wagon” being common in France and Germany.

This was not the first coverage of the Winton Motor Carriage Company in Scientific American.  In July 1897 another article described the first car they built, along with another large photograph:

Scientific American v77 No. 4, July 24 1897, via hathitrust

In the story attached to that photo, it also states that “several firms in the United States are now really in a position to make and deliver motor carriages”.  But even this wasn't the first published photograph of a Winton motor carriage.  The Horseless Age, a monthly journal, published this photo in Novermber 1896, of Winton's earlier first car ever built:

The Horseless Age, v. II No. 1, via Google Books

To clarify, this was the first car built by Alexander Winton, owner of the Winton Bicycle Company, prior to creating the Winton Motor Carriage Company.  The car shown in Scientific American was the first car his new car company.  (It's hard to keep track of all these “firsts” sometimes.)

Let's pause here and try to get some context for this car company.  The Horseless Age where the earliest Winton photo was published was one of several trade magazines in print in 1896 (along with “The Hub” which previously covered traditional carriages and was later renamed “The Automotive Manufacturer”, “The Autocar”, “The Automotor and Horseless Vehicle Journal”, and probably others).  And Winton was just one of what were most likely hundreds of tinkerers that year.  Winton hadn't yet formed his car company in 1896 but there were apparently dozens of car manufacturers in business, including Haynes & Apperson, Duryea Motor Wagon Company, Studebaker Bros. (who as far as I can tell started out making horseshoes), the New York Motor Company, The Reeves Pully Company, P. B. Whitney Motor Carriage Company, and the Elliott Motor Carriage, just to name a few (most names here are from The Horseless Age, volume 2).

An article in The Automotor and Horseless Vehicle Journal from April 1896 mentions that Scientific American had received hundreds of letters in the past year “anxiously inquiring” about buying automobiles, and goes on to say that New York City had already introduced a fleet of horseless taxi cabs.  The article also mentions that a few motor carriages were available for purchase in the United States at this time.

The same magazine (published in London) makes it clear that the U.S. was lagging behind, and indeed The Horseless Age tells us that  England had motor vehicle laws on the books by 1896, and Paris by 1893.  If this seems too early, keep in mind that a steam engine had been attached to a set of wheels to make a vehicle was all the way back in 1769 or so, more than 125 years earlier.  Motorized vehicles were seen zooming up and down city streets fairly regularly by the 1890s, particularly various trucks.  In order to make the transition from trucks to personal vehicles, cars needed refinement, not invention.  They had to become cheaper to buy, cheaper to operate, more reliable, and easier to operate before they were attractive to the average consumer.  The above article from The Automotor and Horseless Vehicle Journal also mentioned that “In its present state of development the horseless carriage can hardly be trusted in the hands of those who have not some acquaintance with machinery”.

One of the companies I left out of the list above, Daimler Motors, was primarily a motor manufacturer, but was also an important manufacturer of motor carriages (as well as boats, which were already commonly motorized).  Take a look at this ad from April 1897:

The Horseless Age, Volume II, No. 6, April 1897, via Google Books

This is not specifically a car ad.  Daimler's biggest business was selling motors to other car companies, and while they did sell motor carriages (at least in Europe), it wasn't their primary business.  But it's certainly an ad in a world where cars were already commonplace in the mind of the public.  By April 1898, Well's Motor Oils was advertising their oils for “Lubricating ENGINES of Autocars, Motor Vans, Cycles, &c.”  You really wouldn't be advertising motor oil for “Autocars” in 1898 if there wasn't a substantial market ready to use your oil for that purpose.

As mentioned, America was running behind Europe at this point, so it shouldn't be a surprise that they had car ads before we did, including this ad in a The Automotor and Horseless Vehicle Journal, (in London) from November 1896:

The Automotor and Horseless Vehicle Journal, November 1896, pg, 84, via Google Books

“A Motor Carriage and Delivery Van can be seen in operation in London by Appointment.”  Is that a car ad, or is it show-and-tell?  It's hard to say for sure if they sell or rent.  But the very next page has one more ad that is not ambiguous:

The Automotor and Horseless Vehicle Journal, Nov 1896. pg. 85 via Google Books

This ad makes it clear: “These carriages are now offered for sale in every variety and description...”, and at the bottom “PRICES FROM £130 UPWARDS.”  This is definitely a car ad, beating the alleged first car ad by nineteen months.  But since this is a London ad, I'll grant my fellow Americans their jingoistic view of the world, and allow that it doesn't count if you assume an implied “American” in the claim for that other ad.

There's definitely other advertisements from specific car companies that beat the original claimed record, including The International Motor Car Company by March 1897, and J. & C. Stirling, from August 1897, as well as ads for “Electric Motor Cars” by April 1897.

So does simple jingoism explain away the claim of that first ad as being the first ad?  I think there's a bit more going on actually.  It seems to me that we're attracted to things that reinforce simplistic views of the past.

The text “Dispense with a Horse” in that ad conforms perfectly to our view of how provincial things were “back in the old days”.  And it conforms with our view that inventions appeared on the earth, complete and ready to go, and changed the world overnight.  As if everybody used horses until one day somebody woke up and said “hey, wouldn't it be easier if we drove automobiles instead”?

The mention of the horse also reinforces the more quaint name “horseless carriage”, and while that was a common name then, it was not the most popular name.  Going by patent searches (via Google), “motor car” turns up the most patents in the 19th century, followed by “automobile”, then “motor carriage”, with “horseless carriage” coming in last out of these four choices.

This provincial feel is reinforced by the coarseness of the image.  It's such a small engraving that blown up it looks a lot like a woodcut, adding to that old-time feel.  Even though photographs were common in print (and this was based on a photograph), this image makes for a more appealing portrayal of our past.  And it doesn't hurt that many versions of this image are somewhat yellowed;   it's unclear if this is the original image capture, or if the image has been yellowed after the fact to ad to the ambiance.

So what?  It's not the first car ad, but it's close enough, right?

As I learn more about how history works, I struggle with the very meaning and purpose of it.  Is it worth the effort to pedantically correct mistakes in our history?  History can never be made whole again—it will always be an approximation.  So if the facts are wrong, does it matter if the essence of the story is still right?  This is dangerous territory, coming very close to saying that the ends justify the historical means.  Can we really trust a simplified version of history to deliver the right conclusion?

In the case of the first car ad, it probably doesn't change things so much.  But in other cases histories which are comforting but false can lead people to fundamentally flawed decisions in the present.

(See, I couldn't even write this trivial bit of history without putting a moral at the end of the story.)

The Lost Key of QWERTY

So I was on twitter last month when Marcin Wichary asked “Any ideas on what this key/glyph was for in the early Sholes Glidden typewriter?”

This image is from a U.S. patent, applied for before the typewriter went to market, but it was definitely there on the first models.

Cropped from an image of the first typewriter model.

And as he pointed out in his tweet, the key produced a typed character that matched the key:

We also know because Mark Twain's daughter was kind enough to type it for us, in some gibberish at the head of a letter he typed for his brother, and leave us with another sample:

These samples eliminate some suggestions that the key served some mechanical purpose, like advancing the paper, or as a shift key (which the first model lacked, as it could only type capitals).

The Sholes and Glidden typewriter (sometimes called the Remington No. 1) was the first successful typewriter ever brought to market (in 1873), and the forerunner of most other successful typewriters.  The unidentified key was, as far as I can tell, on this model and only this model.  It was gone on the Remington No. 2 introduced in 1878, never to appear again (in this form), and as far as I know never found on competitors either.

So what the heck is it?  One option is to work the problem from the modern end, and see what's in Unicode.  We find four characters that look like this:

• ⁝ - U+205D tricolon
• ⋮ - U+22EE vertical ellipsis
• ⫶ - U+2AF6 triple colon operator
• ︙- U+FE19 presentation form for vertical horizontal ellipsis

Some of these are a bit hard to parse.  The vertical ellipsis makes perfect sense, as it is used to show several rows of ommitted information.  But there's nothing in 19th century typography about vertical ellipses, and I haven't even found them in use yet.  Besides, with such a limited keyboard, lacking so many basic characters, why provide this when the colon could serve a similar purpose?  Even if this existed back then, I don't think this was the purpose on this keyboard.  “Tricolon” appears in 19th century sources as a name for one type of verse structure found in the bible, so that isn't so helpful.  “Triple colon operator” unsurprisingly turns up many pages of medical sources, but I found nothing about symbols, characters, etc.

So really Unicode was no help to me.

Next, Marcin Wichary found this character in On the Prehistory of QWERTY.

This is a paper I know well, and have many problems with.  It promotes a new theory that the QWERTY keyboard layout is based on preventing transcription errors by telegraph operators receiving messages.  Given that I've pretty much proven that typebar jams were the primary design goal of QWERTY, we already have problems.  I've promised to write a more thorough debunking of their claims, but it's still on my to-do list.

At any rate, this new claim links to that idea.  The idea here is that telegraph operators needed a way to transcribe the telegraph code for “new paragraph”.  I have confirmed that this telegraph code does predate the typewriter, dating back at least to 1854, but there's still a big problem with this notion.  Our mystery symbol was clearly intended to be typed, however the person transcribing incoming telegrams could just make a new paragraph on receipt of that code, rather than typing a special character.  There appears to be no reason to ever put the symbol on paper.  It's also problematic because it is an incredibly usage-specific character, on a machine that didn't even include parentheses.

I did find a vaguely similar usage of a vaguely similar symbol, in the International Telegraph Code No. 2, which was approved by the CCITT back in 1932.  Many sources use a particular symbol for the line feed character, which looks somewhat like our mysterious three dots, as seen in this document:

But again we have the same problem - there's no reason to ever use this symbol in print, outside of ITA2 documentation.  And we already know the key does not produce a line feed.

A few days ago, the Shady Characters blog picked up this story and ran with the paragraph separator idea that I tend to discredit.  This was followed by a great discussion in the comments section of all sorts of different ideas on the origin, but perhaps the most interesting tidbit came in when a typewriter expert mentioned that his Remington No. 1 produced a slash “/” when this key was typed, rather than three vertical periods.  He referred to it as a “virgule”, which left me a bit confused, as I have Google Books permanently locked in to the 19th century, and the meaning of virgule during that time was just the french word for comma.

So virgule was a dead end, but then Keith Houston who runs Shady Characters (and published a book by the same name), added that it's also called a “solidus”, which lead me to find out that it was even more commonly called a “shilling mark”, and was used in currency, where you'd write “5/” for something that cost 5 shillings, and 5/8 for five shillings and 8 pence (although if something was 3 pounds, 5 shillings, and 8 pence, you wouldn't use / in that case you'd write “£3 5s 8d”, or “£3,,5,,8” or a few other choices).  This was interesting, but it hardly explained why typewriter inventor Christopher Latham Sholes and his cohorts would put a shilling mark on a keyboard that doesn't even have a dollar sign (although you can make a dollar sign by combining S and I).  British currency did last a long time into the 19th century in the United States, but it was certainly not the dominant currency by the time Sholes designed the typewriter.

Moreover, three dots are not a shilling mark, so clearly this was not its initial intent.  But the original three dot pattern did seem to have some kind of relationship with the shilling mark, so I kept digging.

Then I found something really interesting:

The American Bookmaker, September 1887

So while I'm maintaining that a character for the end of the line or paragraph is never used, here's a source saying you need a character for that.  The context for this is the formatting of bibliographies.  In the 19th century books tended to have ridiculously long titles, and bibliographies tended to list the entire contents of the title page.  To do this, multiple lines were joined together, usually (according to this source) with “sidewise” dashes, but in this case also with slashes.  (There's also an implicit reference to letter-cutting, the practice of carving down sorts of various letters to create your own symbols.)

The text above is describing a book called Bibliotecha Hamiltonia, published in 1886:

The slash is being used as a representation for line separators.  I've seen similar usage for condensing a few lines of a poem into a single line.  So here in 1886, we have slash meaning the same thing that three vertical dashes meant 45 years later in the 1930s.  Well, kind of coincidental, but not that exciting.  But there was also this note that the “inclined strokes” were used in place of the more common “dashes turned sidewise”.  Like a vertical bar, maybe?  And so, I was able to find in fairly short order, a vertical bar used like the slashes used in the above bibliography.

A Century of Printing / The Issues of the Press in Pennsylvania 1685-1784, 1886

But in another entry in the same book, I found something else:

There it is.

This is the same three dot symbol used by Sholes on his typewriter, in a context where it is used identically to a slash, also used on the same typewriter by the same key.  In this case, the three dots (which they call here “dotted lines”) are used to show an alternate set of line breaks.  This seems to be a less common usage than the vertical line for this purpose.

This leads me to the following working theory.  Sholes, or one of his testers, wanted a vertical bar character on the typewriter for situations like this one, with a bibliography.  It could be useful for borders and other things too.  But the typography of that first typewriter was stone simple.  It was a sans serif font, and the letter “I” was already a vertical bar.  Given that Sholes doubled up “1” and “I”, there's no point in adding a relatively obscure symbol that was identical.  To be useful it would have to look different than an “I”.  So Sholes simply used an existing alternate form.  Later, when it turned out to be less useful, it was changed to a slash which carried the same function, but could also be used to write fractions, and the percent sign, and to double up with “c” to make “¢”, as well as a number of abbreviations common in that era that used a slash.

It's not a perfect theory.  I have no smoking gun, and I still find some issues with this theory.  But right now, its the best thing we've got that (now) has actual evidence behind it.

The biggest problem is that this is still a relatively obscure usage.  Yes, Sholes was a trained compositor, and would likely have been familiar with all of these symbols, but you'd think he'd also realize these are relatively obscure symbols.  The first keyboard had no at sign, number sign, no parenthesis or brackets, no equal sign, no asterisk, no percent.  This usage doesn't seem to justify this key.  Perhaps there are other usages for the same group of symbols?

One of the big contributors to the development of the typewriter was James O. Clephane, a court reporter who became their best product tester and critic.  His testing lead to a rapid series of changes in design to make the machine more reliable and easier to use.  Perhaps this was a common and essential symbol used in court reporting?  I've searched a bit but come up empty, but it's definitely worth pursuing further.

Still this is all very intriguing.  Is that 1930s document related to this?  It seems to carry the exact same meaning.  And it's only 55 years later, still within the memory of some.   If I'm on the right track, this mysterious key could have lead to the inclusion of slash on the keyboard and in ASCII, and as well as both the vertical bar, and the broken vertical bar (which was created for ASCII in the 1960s to avoid confusion with the mathematical “or” operator).

There's even a tie-in with the pound sign.  In part two of my examination of that character, Britain on Hash, one of the possible-but-unlikely origins of the name hash for the pound sign was a practice of using slashes and dashes in a new piece of computer technology called KWIC indexing, where the separator usage of slashes and dashes seems very much related to this old usage.

The “lost” key might not be lost at all, just changed over time.

Of course we have to be realistic.  This is all just guesswork — the ramblings of a madman.  The only thing we can really say for sure, which is still a step forward, is that a symbol just like the original three dot typewriter symbol was used to indicate line separations in typography, and that this symbol was replaced on the keyboard with a slash, which was also used in typography for the same purpose.

The truth may still be out there.  There is supposedly an original catalogue that came out with the first Remington typewriter, a multipage pamphlet called The Type-Writer: A Machine to Supercede the Pen, which may well describe this key and its purpose.  So far all I've been able to find are simple single-page ads with that text.

Update: November 2019

In April of this year additional information came to light from Eric Fischer in a Twitter thread (started by Mr. Wichary) suggesting that the symbol was a substitute for parenthesis and braces.  The source provided is from 1887, by which time I would have thought the key had already been replace.  Wichary says he found another source that backs this up, but also suggest that the tricolon could be used as a combining character with S to make a dollar sign—though he rightly questions why you wouldn't just combine with the "I" as was common known practice, and which was even mentioned in his source.

I had found an early hint that this key was in fact meant to be used a both left and right parenthesis, and even though I tweeted about it at the time, didn't think it was likely enough to merit a mention in this article.  Yet another lesson that you should always include everything.

The Horror of France

Just a quick update with some good passages I've come across.

The first, is from Bookbinding and Book Production, Volume 50, page 76, from 1949.  (And just to be completely clear here, since there is great confusion over both our printing history, and the term “French Spacing”, the change discussed here is the move from wide sentence spacing to narrow sentence spacing.)

French Spacing? Horrors!

Reports that “French Spacing” might become more widely accepted on account of savings in money and time has caused at least one designer to remark “Preposterous.” Another threatened “I'll shoot any compositor who does,” and another claimed that someone was just trying to be different—“a brand of lunacy.”  Only one voice spoke in favor of such a trend.

 “French Spacing” or word spacing between sentences is one of the recent typographical innovations at the Waverly Press, Baltimore, Md.   With the advent of shoulder-spaced type, the change became an economic issue as well as a typographic one. Only because accountants speak louder than artists, the change was made.  The advantage is that every time the end of a sentence is reached, one character is cast instead of two.  The actual saving in caster strokes while not heavy percentage-wise is very high numerically with the millions of caster strokes made each week.

I tend to say that the transition period from wide to narrow sentence spacing is from about 1930 to 1950, but it's clear that there were some holdouts even at the end there, and that they felt as strongly about this as modern one-spacers.

I disagree with their simplistic analysis of the cost.  Simply the cost of typing the key was not a major issue.  (For that matter, neither was savings in paper, although in both cases, that was an era of particularly overzealous bean-counting.)  As mentioned in other articles here, it was a constellation of factors: it saved in error corrections (once the habit was relearned); it reduced the rivers that were a particular problem of the Linotype; it prevented machine downtime when typists were used on teletypesetting systems; etc.

They are one of the very few sources I have found, ever (before like the 1970s) that use the term “French Spacing”, which as I have said elsewhere has undergone a reversal of meaning (and an incomplete one, meaning that modern sources contradict each other).  French spacing originally was word spacing between sentences, versus the wide spacing we used.  Some modern sources completely reverse this.

I have a theory that we'd never even know about “French Spacing” if it wasn't for the \frenchspacing operator in Donald Knuth's TeX typesetting language.  But we do know about it, and as I've said in my other article, the change was probably inevitable.  But it seems to have started (in spirit at least) sooner than I thought.

In 1960, an article was found in The Inland Printer and American Lithographer (or some vaguely similar name to that which seemed to change repeatedly over the years), volume 145, from 1960, as well as Typo Graphic (some volume or another), also from 1960.  It was found in a “Question & Answer” section.

Q. What is French spacing?
A. There is no reference to this term in any textbook on printing or in any glossary. In searching for a precise answer I therefore turned to a number of the leading typographers in New York. To my surprise, a good many of them had never heard the phrase. However, I was finally able to come up with the following:
    French spacing is tight spacing, with equal word spacing throughout a line, i.e., no extra space after a period, colon, etc. The purpose is not only to create a tighter looking, evenly colored page, but, more important, to avoid rivers. In some ad shops, French spacing is understood to mean optically equal word spacing. As to the “French” part of the term, this style has nothing to do with France as verified by several French cultural societies and printers. The word was evidently used because anything “French” was considered to be du haut style.

This confirms what I suspected, that this was a rarely used term.  But also, at the end there, they deny that French Spacing had anything to do with French practices, instead assuming that word spacing was called French as a sort of compliment.  Step one in the rewriting of the history.  Step two of course, was to recognize that nothing truly American could ever be called French, and we have a full reversal.

I also assume the claim that they actually contacted French printers was either a lie, or they contacted someone who was clueless.  As far as I'm aware, France never used wide spacing, although I admit I haven't researched the subject as thoroughly.

The first quote also opens up another avenue of exploration: I have to go and figure out what the “shoulder-spaced type” in that first quote means.  Some hints I've seen point to yet more technological issues that both benefitted and hampered the printing process.

Summary of arguments in favor of two spaces between sentences

Functional
• The period is an overloaded piece of punctuation with multiple meanings. Periods alone are ambiguous or misleading sentence termination.
• Two spaces allows for vastly simpler and more accurate machine interpretation, including translations and text-to-speech.
• Ironically, this machine-interpretation issue allows for simple technology to allow the reader to control visual layout to their aesthetic preference. One space between sentences forces readers into the writer's preference.
• Wide spacing has been shown to be better for many new readers, and for some learning issues.
• In my opinion, wide spacing is clearly superior for skimming and scanning, and for finding your place again.

Aesthetics
• “Everybody does it that way now” is just fashion, not a design law carved in stone. And it ignores the history of how we got here.
• The aesthetic argument is 100% undercut by the functional argument of two spaces allowing for reader-controlled display.

• Human beings found wide sentence spacing preferable for four hundred years.

History
• All the “just-so” stories about sentence spacing you've ever heard (typewriters, monospaced fonts, etc.) are easily disproved.
• The only compelling and supported theory on how we lost wide sentence spacing is that the technology couldn't handle it.

Summary
Ultimately, we provide more information with two spaces between sentences. This trumps everything, particularly the aesthetic argument, where one space is dictatorship and two spaces can be reader-centered.