Chapter 18: Paper and the Printing Press
Era span: ~100 CE paper → 1450s press · Difficulty: mid
Requires: Ch 11, Ch 15, Ch 16
Unlocks: Ch 20, Ch 36
Printing is the meta-invention: it accelerates the invention of everything else by collapsing the cost of copying exact information — diagrams included, which speech cannot carry. A civilization that prints cheaply runs its technology tree in decades instead of centuries.
18.1 Paper
Paper is a random mat of cellulose fibers drained on a screen:
- Pulp the fiber: shred rags (linen/hemp/cotton — best early feedstock) or bark; beat in water with stampers (Ch 16 water-hammers) until fibers separate into slurry.
- Dip the mould: a wire-cloth screen in a wooden frame lifts a thin layer of slurry; shake side-to-side to interlace fibers ("vatman" and "coucher" as a two-person team).
- Press and dry: couche wet sheets onto felt blankets, stack-press out water, then hang-dry.
- Size (surface-treat) with dilute gelatin or starch so ink sits on the surface instead of feathering.
Wood-pulp note: mechanical wood pulp is cheap but lignin-rich and ACIDIC — books self-destruct in decades (Dead end avoided: building your whole archive on high-lignin newsprint; use rag paper for anything permanent, or chemical pulps once alkali processes exist via Ch 21).
| Feedstock | Prep | Sheet quality | Verdict |
|---|---|---|---|
| Linen/hemp rags | Retted, beaten 4–8 h | Strong, archival, takes sizing well | Default for permanent books |
| Cotton rags | Shorter beating | Soft, absorbent | Good; slightly weaker than linen |
| Mulberry / kozo bark | Steamed, stripped, beaten | Long fibers, tough, thin | Best for large sheets; Asian standard |
| Straw + lime | Cooked alkaline, beaten | Decent, yellows slowly | Volume paper once alkali exists |
| Groundwood (mechanical) | Ground wet, no chemistry | Cheap, lignin-rich, browns | Newspapers only — never archives |
Vat-team rhythm: the vatman dips and shakes (~10 s per sheet), the coucher transfers to felt (~8 s), the layman stacks. A practiced trio lifts 2,000–3,000 sheets/day. Drying loft capacity — lines, air flow, frost-free space — limits output more often than the vat does; size the loft before hiring the second team.
RAGS → sort/wash → STAMPERS (water-powered) → slurry vat
│
MOULD dip + shake → COUCH to felt → STACK PRESS
│ │
└→ HANG-DRY → SIZE → dry again → calender
18.2 Ink
Water-based carbon inks bead on metal type. Printing needs a viscous varnish ink: boil linseed oil (or walnut) to varnish thickness, stir in lampblack until glossy black. Varnish temperature controls tack; too stiff, it pulls type out of form; too loose, it squashes.
Reference batch: ~2 L linseed oil boiled outdoors (fire risk — it foams and can ignite) until a cooled drop strings ~2 cm between fingers; stir in ~400–600 g lampblack plus a spoon of rosin varnish. Grind on a slab with a muller until no grit shows between teeth. Store under water film in sealed pots. Test-tack rule: dabbed type should print solid black with crisp edges at one firm pull — halo squeeze means too loose, grey mottle means too stiff or under-inked.
18.3 Movable Type Systematics
Safety warning: type metal and inks poison slowly — molten lead fumes and dust accumulate, antimony fumes sicken, and oil-ink solvents flash near dryers. Melt and cast only ventilated with fume capture, keep food and drink from benches, wash hands and cover cuts before meals, and store solvents from heat with extinguishers staged.
Gutenberg's real invention was the SYSTEM, not the idea of movable type:
- Punch: cut a steel letterface hard.
- Matrix: strike the punch into soft copper — the mold cavity.
- Caster: an adjustable hand mold clamps the matrix while molten alloy pours in; every letter body comes out identical height.
- Alloy: lead + antimony + tin (foundry type roughly 60 Pb / 25 Sb / 12–15 Sn; Gutenberg-era and later machine-casting alloys were leaner in both). Antimony expands slightly on cooling — crisp letters, no shrinkage cavities. Pure lead shrinks and rounds off.
Type height standardization (all foundries same dimension) lets any printer buy any founder's type — network effects from day one. Cast in quantity: lowercase/uppercase/punctuation in proportion to language frequency; compositors pick from two-compartment cases.
Sort quantities per 100 kg fount (Latin text guide): e ~12 %, t/a/o/i/n ~7–9 % each, full lower-case ~60 %, capitals ~8 %, figures/punctuation/spaces the rest. Running short of e stops the press as surely as running out of paper — cast frequency-weighted, not alphabetically.
18.4 The Press
A wine-style screw press adapted: flat bed holds the locked forme (type face-up), paper held in a hinged "frisket/tympan" sandwich folds down, screw drives a platen pressing one page area per pull. Two trained men run ~200–250 sheets/hour single-sided. Compose → lock (quoins wedged against furniture) → ink with leather dabbers → print → distribute type back to cases. Page imposition (arranging pages so folded sheets read correctly) is bookkeeping-level cleverness that halves binding labor.
Key threshold: one press + crew produces more identical pages per year than a scriptorium of hundreds — at falling marginal cost. The cost per copied page fell by orders of magnitude as presses spread; expect a similar collapse here.
Press-crew SOP (two pullers): (1) plane the forme (tap the type level with a planer block) and wash it with lye; (2) dab ink twice, perpendicular passes; (3) fold frisket/tympan, crank bed under platen; (4) single firm bar pull — no double-strike (it ghosts); (5) crank out, lift tympan, hang sheet; (6) every 50 sheets, proof against copy. Wash forme nightly; pick out kerned sorts with an awl, never a knife edge.
18.5 What Mass Printing Changes
- Error rates: copies are identical; errata sheets fix all copies at once. Knowledge stops decaying in transmission — the silent killer of pre-print craft secrets.
- Standardized diagrams and tables circulate: geometry figures, machine drawings, anatomical plates, star charts (Ch 20 depends on exactly this).
- Literacy pays: cheap primers make universal schooling rational (Ch 47); vernacular languages standardize.
- Critique scales: identical texts can be compared, cited, argued over page-by-page — the precondition for peer review culture.
- News networks: periodicals synchronize markets and politics across distances.
Jump: skip woodblock-only era for books (keep blocks where they shine: images) — go straight to full movable-type systems using §18.3's punch/matrix/caster trio. Also skip hand-press refinements directly toward iron frame presses (lever-driven; a larger platen prints a full forme in one pull with far less effort) once foundries exist.
| Stage | Machine | Output | When |
|---|---|---|---|
| Hand press, wood frame | Common press, 2 men | ~200–250 impressions/h | Day one |
| Iron hand press | Stanhope/Columbian levers | ~250–300/h, full sheet per pull | Once iron founding exists |
| Cylinder, hand-fed | Koenig-style flatbed | ~1,000–2,000/h | With Ch 23 power nearby |
| Rotary web | Steam-driven reel | 10,000+/h | Full industrial stack |
18.6 Scaling Doctrine
Sequence: one press per literate market town → newspaper weeklies → stereotype casting (whole-page lead plates for reprint runs) → steam-powered cylinder presses (Ch 23) → web-fed rotary monsters printing tens of thousands hourly. Each step rides existing tech; none requires new physics. Printing is infrastructure — fund it like bridges.
Stereotype payoff: cast a papier-mâché flong of a composed page, pour type metal to make a flat plate (or a curved one for rotary presses), print from the plate and distribute the type. Reprints no longer tie up the fount — the book stays "standing" as plates on a shelf.
Pictures, maps, and lithography: type carries words; diagrams need their own processes. Woodcuts print alongside type but hold only coarse lines, so fine maps, instrument scales, and technical plates were cut or etched into copper (intaglio) and printed on a separate press. Lithography (Senefelder, 1796) removed the engraver: draw with greasy ink on a smooth limestone, wet the stone, roll on oily ink, and the ink sticks only to the drawing because water repels it everywhere else. Maps, diagrams, music, and labels then print as fast as they can be drawn. Photographic transfer onto zinc and aluminium plates and the offset press (an inked image passed via a rubber blanket to the paper, early 1900s) made lithography the dominant commercial process, and its light-sensitive coatings are the direct ancestors of photolithography (Ch 35 §35.3).
18.7 What the Incunabula Record Shows
- Gutenberg's 42-line Bible completed ~1454–55 (~180 copies printed; ~49 survive). The Fust–Gutenberg lawsuit of 1455 documents who financed and who lost the original venture — printing began with litigation over money, like most industries.
- Diffusion speed is the dataset that matters: by 1500 — just 45 years after Mainz — presses operated in roughly 270 European towns, producing an estimated 27,000–29,000 distinct editions ("incunabula") totaling millions of copies. Few earlier technologies spread anywhere near that fast.
- Venice's Aldus Manutius invented the portable book market (octavo classics, italic type cut by Griffo) — product design as market creation.
- The Reformation as print event: Luther's 1517 theses circulated across German-speaking Europe within weeks; his 1520–22 tracts sold hundreds of thousands of copies. Whether printing caused the Reformation or merely carried it, historians broadly agree that censorship became far harder to enforce at scale — governance changed permanently.
- Scientific communication institutionalized: the Journal des sçavans and Philosophical Transactions appeared in January and March 1665 respectively — priority registration, citable results, and peer critique begin here (Ch 20's method gets its postal system).
- England's first press: Caxton at Westminster, 1476; Germany's first documented paper mill: Ulman Stromer's Nuremberg mill, 1390, sixty years before the press — paper capacity preceded type capacity, a supply-chain lesson rebuilders should note (§18.1).
18.8 Troubleshooting Matrix
| Symptom | Cause | Fix |
|---|---|---|
| Feathered, fuzzy letters | Unsized paper or too-loose ink | Re-size sheets; stiffen varnish, lighten dabbing |
| Type pulls from forme | Ink too tacky or lockup loose | Thin ink slightly; re-wedge quoins, check furniture |
| Ghost / double image | Bed moved during pull or double-strike | One firm pull; tighten rail gibs |
| Uneven impression (pale corner) | Platen unlevel or forme unlevel | Shim with tissue under forme; level platen screws |
| Rapid type wear, rounded faces | Too-soft alloy (lead-heavy) | Re-cast toward ~60/25/13 Pb/Sb/Sn; cool moulds fully |
| Sheets stick, tear on lift | Over-inked or damp stock | Less ink, more drying time; dust with dry starch |