In front of you are two drawers, called cases. They’re full of letters, split between a case you place higher — an upper case — full of capitals, and a lower case of minuscules, or small, literal “lowercase” letters. In the dim light, you glance at a piece of paper covered with handwritten words and notations and, with a practiced motion, rapidly grab pieces of metal from compartments in each case and stack them successively, left to right, in a metal stick in your hand, forming lines. To justify a line—making the left and right edges fill the width of a column—you recursively add ever-smaller spaces between each word until the line is firmly packed. When the stick fills with lines, you slide its contents onto a metal tray and start to fill the stick again.
You work at this repetitive task for hours on end. If you’re good at your job, you’ve composed what amounts to a few pages of a book by the end of your shift. Exhausted, you stumble out of the asphyxiating gaslit air toward your favorite typesetters’ saloon to gather with union mates and drink heavily. At some point after dawn, you stagger onto a streetcar or train and head home to bed. On waking, you almost immediately set out for work and start all over again, typically arriving an hour or two before your real shift to distribute the type, taking apart the previous day’s work and sorting it back into those cases’ cubbyholes.
This was the lot of the typesetter or compositor in the mid-1800s, a profession that had become more refined since its invention as a trade by Gutenberg in the mid-1400s, but which largely involved the same tasks. Gutenberg’s key invention that sparked the printing revolution, as I described from a couple angles in the first two parts of this series, was a type mold that allowed the creation of individual, reusable, “movable” pieces of metal type.
The difficulty in creating original designs in hardened steel, which ultimately produced metal type for printing, was removed as a gating factor in the late 1800s. Type foundries suddenly could produce more designs, more sizes and styles of type, and more of the matrices used to cast type in metal.
Yet typesetting mostly stood still: It was an intensely skilled, intensely manual task. As the processes to print pages increased in speed by hundred- to thousandfold, newspapers, magazines, book publishers, and other shops could increase the variety of content they published only by hiring more compositors. It didn’t scale.
Typesetters had reached the limits of human frailty in relation to the efficiency of machines. They couldn’t move any faster. The only way to produce more pages was to hire more people, but the profession had a multiyear apprenticeship to learn all the tricks of the trade and build up speed.
These workers had every reason to be as fast as possible, as they were paid piecework until quite late in the handset era. As a mostly unionized profession, they were paid by groups of 1,000 “ems,” a measurement that once meant the width of the capital letter M in a font but was standardized to the type size of the font in points.
In the 1800s, there were roughly 72 points to an inch, so 1,000 ems in 12-point type would mean 12,000 points or about 14 feet of type. That’s about a single page in an average-size hardcover book and a portion of a broadsheet newspaper page. The records of the time indicate a good compositor could set on average about 700 ems an hour across a 10-hour shift. (The one or two hours spent distributing type wasn’t compensated, as it was considered part of the cost of piecework.)
At a fast pace, typesetters could reach 1,500 ems an hour. In races — yes, there were typesetter races — the fastest, known as “swifts” or “fire-eaters,” could exceed 2,000 ems. Some publishers disliked swifts: They ruined morale for more ordinary setters.
Typesetters were also highly skilled labor, earning more than a blacksmith or machinist, according to 1870 records, translating into the equivalent current buying power of somewhere in the $75 to $100 per hour range. The schedule of prices set by the union for varying sizes of types, foreign languages, and other variables was extensive and elaborate, designed to match time-consuming or difficult tasks to a higher wage to compensate for hours or expertise.
Despite strong unions, working conditions were terrible, full of lead dust and regularly spread consumption and other diseases. The average lifespan of a typesetter in 1850 was 28 years. All those factors led to chronic shortages of compositors.
Skill and mortality restricted publishers’ ability to fill the demanding mental maws of increasingly literate Americans and Europeans. But what if a machine could take the expertise out of a worker’s hand and make them more of an interchangeable cog?
Ah, the constant dream of capitalists, but one that would benefit typesetters as well, with more regular wages and hours, and, as it turned out, much longer lives.
A Rube Goldberg Machine That Predated Rube Goldberg
Many inventors tried to create a button- or lever-driven device that would allow inexperienced workers to set type. Most of these used a process of mechanically selecting individual metal characters or funneling them from compartments and dropping them into a type stick. Because women were largely excluded from typographical unions, they were often hired — at much lower wages — to operate these wonky machines, which used primitive keyboards and constantly locked up or broke.
A fully automated system that would allow high-speed keyboarding to set type seemed complex beyond reduction, partly due to the same sorts of limitations in metallurgy and manufacture that held up aspects of Charles Babbage’s various computational engines. Those fields finally matured enough that Ottmar Mergenthaler could develop and perfect the thousands of parts that allowed his Linotype typesetter to function in the 1880s. (Samuel Clemens famously went bankrupt in part due to his investment in the Paige Compositor, a system that worked but was more temperamental in operation than the Linotype.)
The Linotype could set, literally, a “line o’ type.” It had several unique aspects that dramatically sped up typesetting by replacing human labor with mechanized action. Instead of using type cast by a foundry as individual characters, the Linotype cast its type on demand. The machines had large dispensing magazines, like the innards of a soda-can machine, that contained matrices for casting type. Linotypes could eventually hold or switch between multiple fonts at once, allowing a shift between Roman and italic type or among entirely different typefaces.
As a compositor tapped away, the Linotype released matrices that corresponded to each character from the magazine, which slid into place in a casting section to the left of the keyboard, just as if they were being set in a handheld typestick. When an operator pressed the spacebar, the machine inserted wedge-like spacebands between words. When a line was complete, the operator triggered justification, which pushed the spaceband wedges forward until the line was firm.
The next step you might not believe if you’ve never heard of this before: Boiling lead, kept burbling in the belly of the machine, would shoot out and fill the matrices! This produces a solid and neatly cast line, which was pushed by the machine into a galley. If something went wrong, that boiling lead could also shoot out and hit the typesetter. This was known as a “squirt.” From a 1915 Illinois factory inspection report:
Linotype and Monotype machine operators frequently are burned. Linotype machines are dangerous in this respect, because the operator is seated in front of the machine and is likely to have the metal splashed into his face.
With proper precautions, like metal shields and operators listening carefully for the telltale sound of an impending squirt, typesetters could — largely — avoid burns. But few made it through a career unscathed. Carl Montford, an 80-year-old retired Boeing engineer who acts as a sort of godfather to the Seattle letterpress community, recently told me about his time as a 13-year-old managing the lead, including melting down the cast type after it had been printed at a small Midwestern newspaper. This job, often given to boys even younger than 13, was known as the “printer’s devil.”
This brings us back around to Part 2 of this series. The Linotype would have been impossible without Linn Boyd Benton’s pantograph and related machines, because the hot-metal systems required huge numbers of matrices. In a 1919 publication, the Mergenthaler Linotype Company said it produced 250,000 unique characters for its machines worldwide and stamped out 1 million matrices a week.
Without Benton, the Linotype would have stalled, as its manufacturer required a volume of production that had to be matched by the availability of matrices to run the systems in the field.
Cheaper, Faster, Better, Simpler
The Linotype was clever enough at setting type, making it possible for a trained operator to routinely compose around 5,000 ems an hour and the most skilled as many as 10,000 ems an hour. But, you might ask, what about sorting all those matrices back into their compartments in the magazine?
This was yet another facet of the Linotype’s cleverness. The bottom of each matrix had a V-shaped channel full of notches. After a line was set, the operator triggered distribution, which carried each matrix back to the top of the magazine, and then it sorted through those notches into the correct compartment.
The Linotype also helped with a related problem: the cost of type. In 1917, when handset type still abounded for a lot of uses that didn’t involve high-speed production, a trade publication put the price of standard type at about $300 to $600 per font drawer in today’s dollars. A shop needed hundreds of drawers to have enough sorts (individual characters) and enough variety to meet their needs. That was a lot of capital. And the type wore down, requiring continuous replenishment.
The Linotype changed that entire equation. A report from Harvard’s business administration school in 1916 showed a savings to printers of nearly two-thirds using a Linotype compared to handsetting. The report didn’t address speed, which was several times that of an experienced hand compositor. The cost per hour was a lot greater, but the labor multiplier much higher.
Typesetters also required less training in arcana and were expected to perform to a more professional, sometimes overbearing standard of office-style work, something still emerging in the late 1800s. I visited a print and design shop with a fully functioning Linotype six years ago and, with a few minutes’ instruction, was able to set right to work despite the unfamiliar keyboard layout.
Switching to hot metal didn’t reduce the demand for typesetters, because the acceleration in producing words faster led to lower prices for periodicals and books, which in turn fed literacy and demand.
Linotype wasn’t the only game in town. It was joined soon after it gained traction in the mid-1880s by the Monotype, a competing system preferred for book typesetting, as it used an even cleverer system to cast lines of type as individual metal sorts using a different matrix approach.
Instead of the danger of lead squirts as one typed, the Monotype system split composing from casting. A compositor would sit at a keyboard that punched holes in a roll of paper tape. That roll would be taken to another part of the shop to a casting machine and “played back,” almost like a player-piano roll, and the perforations would signal which letters to cast in sequence.
Typos could be fixed by changing out cast sorts or by fixing the paper tape, like debugging. The Monotype system presaged the computer era by abstracting and dividing functions, and paper tape was an early storage method used for writing software offline and storing programs for later use.
Linotype, Monotype, Intertype, and others persisted for several decades, improving and maturing, until almost all at once, hot metal disappeared. A new revolution, this time involving photography, swept aside hot metal and cold foundry type alike. Photographic-based typesetting (“phototypesetting”) was to the Linotype as the Linotype was to handset type.
Phototypesetting used photo-development chemicals instead of boiling lead and allowed a single character etched in film to be reproduced nearly infinitely at the cost of only a special kind of photosensitive paper. It killed hot-metal composition almost completely, but it wasn’t the end. It was a John the Baptist: It presaged the rise of digital type, which left phototype dead.
See Linotypes in Action
For more about the Linotype, you can consult Linotype: The Film, a documentary about the machine’s history and the people who ran them, and Farewell — ETAOIN SHRDLU — 1978, a film shot on the last day that the New York Times operated its hot-metal typesetters.