In his final decade, the Shakespeare scholar and bibliographer Charlton Hinman (1911–1977) began to lose his sight. Writing in an obituary published in The Book Collector, Fredson Bowers — Hinman’s friend and onetime academic advisor — observed that the scholar’s near-sightedness was the result of “excessive strain over the course of years and too many hours of concentrated working the Collator” (Bowers 1977, 391).
Working the what?
TheCollator Bowers refers to here is an optical-mechanical device invented by Hinman to compare two copies of the same book. Since Hinman’s death, the collator has taken on the name of its inventor and is now known as the Hinman Collator.
Scholarly work in the humanities is not without its own toolkit of technologies, some ancient (the book wheel, the magnifying lens, the index) and others of more recent invention (topic modeling, digital facsimiles, the card catalog). The printed book is itself a kind of technology, one uniquely suited to the human hand and mind — explaining its persistence in the face of newfangled alternatives. But the strangest of these humanistic technologies, both in appearance and function, is perhaps Hinman’s Collator. With its angular framework, pushbutton switches, and Bakelite nameplate, the collator betrays its origins in postwar academia. Ungainly and poorly balanced on too-small wheels, it resembles a diminutive, vinyl-upholstered ziggurat — a strange mid-century altar onto which countless scholars have placed their bookish offerings. But when in use, the collator’s oddities become purposeful, and the device takes on a less unnerving guise. It is an object by turns disquieting, useful, and endearing.
The Hinman Collator enacts an optical test known as the flicker paradigm — a visual phenomenon by which the brain detects subtle differences between two alternating images. When using the collator, the operator places two copies of the same book on cradle-like platforms underneath lamps that blink in rapid alternation. After bringing the two copies into focus and alignment, the operator peers through a binocular eyepiece, starts the device’s blinking lamps with its knee-activated switch, and sets in motion the collator’s optical effect: in one blink, the operator sees one copy, and in the next, the other. This rapid blinking — A then B, B then A — causes the two copies to converge and superimpose, tricking the eye into seeing two copies as one. Any printed variants — damaged type, alternate spellings, changes in page layout — appear to shift. And the eye’s failure to resolve the two images makes visible even the smallest of discrepancies.
One writer has likened the flicker of these optically disclosed variants to a “dance or wiggle,” though the Hinman Collator offers little in the way of grace or choreography. Charlton Hinman himself admitted that “the machine is at once awesome and a little ludicrous. […] it is ungainly looking at any time; but it presents an especially odd spectacle when in use.” Gesturing again toward the device’s strange aspect, Hinman noted that “Rube Goldberg is inevitably suggested.” The collator’s inner workings bear out the comparison: the device’s on-switch, a dial-like potentiometer that tracks rotational distance to set the speed of its blink, resembles (or even is, it’s hard to tell) an emptied spool of thread. In place of the absent thread is a nylon cord wrapped around the spool’s aluminum barrel and tied to a loosely coiled spring that lends the switch its squeaky recoil. The device also contains a kind of analog memory bank, again recalling the inscrutable mechanics of a Rube Goldberg machine: behind a cabinet door near the user’s knee is a counter labeled in uppercase letters, “HOURS”. The Hinman Collator held in Carnegie Mellon University’s Posner Center, where I work, records 5,118 hours of on-time, or more than six months of continuous collating.
Machine vision is now the stuff of algorithms and computer-speak, which makes the Hinman Collator’s cog-and-wheel plainness all the more endearing: it is an object of nostalgia as much as it is one of scholarly pragmatism. Even the heavy suffix of the device’s name recalls an earlier and simpler relationship with technology: the agent noun, collator, personifies the machine, evoking a thinking, breathing collaborator instead of an inert instrument or tool. It has a sort of inelegant charisma that immaterial algorithms and code lack.
But what is meant when one claims to collate? What action does a collator perform? The etymology of collate is tangled, deriving from the past participle of the Latin verb conferre (“bring together, collect, gather, contribute, connect”). But in English — and particularly in the overlapping fields of rare books and bibliographical analysis — it has taken on a narrower meaning. To collate is to compare, but more strictly it means to place two copies of a book side-by-side in order to reveal something about the obscure practices of the premodern print shop. Collate can also imply a kind of confirmation: a process by which the completeness of a book is verified and recorded. This is the meaning of the word most often intended when it’s used by antiquarian booksellers and auction houses, whose interest in completeness understandably takes precedence over the scrutiny of minute variants in the text.
Linked here is a video that approximates the stereoptical effect of the Hinman Collator’s blink. It shows a passage in Act Four of Shakespeare’s Othello in two different printings. In the later printing, “set sining” is revised to “sat singing” in the first line. Between the two versions, or “states”, the typesetter has amended the phrase, replacing the e in set with an a and adding a g between the n and i in sining. Both readings, it turns out, are wrong: pointing to the frequent unreliability of premodern printing, the line should read, “The poore Soule sat sighing, by a Sicamour [sic] tree,” as is clear from the context of Desdemona’s lament.
By enabling the detection of variants like this one, the Hinman Collator allows scholars to judge which printing of a line is closest to what Shakespeare actually wrote. In this example of an error in Othello, Hinman’s machine opens a window onto a moment of orthographic and interpretive failure; its rapid blinking permits us to witness a seventeenth-century typesetter at work, attempting to make sense of an ambiguous or smudged line in his manuscript copy.
Meanwhile, in the variant states of Shakespeare’s engraved portrait — also shown in the video linked above — we find more evidence of alteration. The changes revealed by the collator’s blink represent the engraver’s effort to modify the image to more accurately depict Shakespeare both “after the life” and after life (Shakespeare had been dead seven years when the plate was made; the image, more ghostly remnant than embodied face, was evidently made without a model). The two states of the portrait are distinguished by lines added to darken the shadow cast by Shakespeare’s head onto his vast collar. This cross-hatching lends the later state more depth and color. But Shakespeare’s beard is also made thicker in the later state, making it appear in the video as though the Bard’s face flutters from clean-shaven to stubbled and back again.
For all its clockwork strangeness, the story of the Hinman Collator’s invention is less ludicrous than it is practical and commercial. After a prototype emerged in the 1950s (Hinman was an inveterate tinkerer and had been a Navy cryptanalyst assigned to a code-breaking unit during the Second World War), Hinman contracted with Arthur M. Johnson, a retired Navy engineer, to fabricate collators to order. The collator was briefly in demand, viewed by curators and influential professors of literature as a required piece of kit for research with rare books and prints. Seeing its potential — Hinman had published a string of articles hinting at the collator’s astonishing alacrity — literary scholars, bibliographers, museum curators, and textual editors who had access to the device set about tracking and explaining variants and typographic anomalies in early printed books. The collator allowed these scholars to attain previously unimaginable accuracy and efficiency in their work.
Hinman used the Collator in his own exhaustive analysis of the 1623 First Folio of Shakespeare’s plays — a project that led to the publication of a two-volume study, The Printing and Proofreading of the First Folio of Shakespeare (Oxford, 1963). The story of Hinman’s optical-mechanical analysis of the Folio is a long one: he eventually examined fifty-five copies of the Folio over many months with the help of his collator at the Folger Shakespeare Library in Washington, D.C. Beyond cataloging and tracing 600-odd distinctive types through the Folio’s 900 pages, his findings revealed a spate of previously unremarked details about the Folio’s printing. His method meanwhile reshaped the field of analytical bibliography.
It’s difficult to say precisely what the collator signaled to those not involved in its scholarly use. An article that appeared in the New York Times in 1950 hints at both the era’s breathless embrace of a machine future and amusement at the lengths to which tweedy scholars might go for ever diminishing returns: “Reading Shakespeare by Machine,” the headline reads.
But Hinman’s collator did find a kind of traction and wandered beyond the world of Shakespearean textual scholarship and typographical detective work. A report by Elizabeth Mongan — Curator of Graphic Arts at the National Gallery of Art — submitted in support of a grant proposal to fund the collator’s manufacture notes that her “particular interest was to test [the machine’s] usefulness in respect to print,” including “engravings and etching by Schongauer, Durer and Rembrandt.” Having examined prints by each of these artists (unnervingly, Mongan used the phrase “in the machine”, implying a kind of mechanical ingestion or assimilation), she marveled at its “remarkable possibilities”, reporting that it “disclosed undescribed differences of lines, cross hatching, etc.” The device, Mongan concluded, represented a “vast time saver in that it greatly speeds up the arduous work of meticulous comparison.”
Hinman’s instrument-aided method, which he called “mechanized collation”, proved influential: though Hinman’s was the first, several optical-mechanical collators followed, including the Lindstrand Comparator, Hailey’s Comet, and the McCleod Portable Comparator — all named for their slightly madcap inventors. Even more recently, PocketHinman, a smartphone app, brings the Hinman’s blink to the scale of a handheld screen, allowing users to overlay digital photographs of variant states of a text for comparison. Each of these solutions is meant to identify a common feature of early modern presswork: error.
In seventeenth-century print shops, including the one that produced Shakespeare’s First Folio, mistakes in typesetting were corrected during printing. But rather than discard the early, error-riddled sheets, uncorrected copies printed before proofreading were bound up and sold, meaning that editions of early printed books almost always comprise a jumble of sheets, some in their corrected state, some not. Determining what Shakespeare actually wrote, Hinman realized, would thus require examining several dozen copies of the First Folio, the only reliable witness for nearly half of Shakespeare’s plays. As Hinman explained in 1953, “we cannot hope for the best possible edition of the text of Shakespeare until all of the variants in the First Folio […] have been discovered and recorded.” The Hinman Collator made this painstaking search for variants practicable.
What fascinates here, though, rather than the device’s historical applications, is the thing itself. According to a census published in 2002, there are just over forty-one Hinman Collators that improbably survive. Some of these are in various states of disrepair and decrepitude. Still, this is an unexpected number — the survival of forty-one collators is a reminder of the brute endurance of these machines: only fifty-nine were manufactured, the last sold in 1979. The Collator at Carnegie Mellon came to Pittsburgh by a circuitous route. Long kept in storage at the University of Wisconsin’s surplus warehouse in Milwaukee, it was removed from the Shakespeare Research Collection in the University of Wisconsin’s Golda Meir Library — its home since the University purchased it from Arthur M. Johnson in 1964. Brought to my attention by a cherished mentor and professor, I acquired it for the cost of freight, and after some logistical wrangling, it arrived on the back of an open trailer, shrouded in industrial-strength cling film.
Made obsolete by algorithms that detect printed variants at superhuman speeds, the Hinman Collator is no longer a useful object. Or, more accurately, it is no longer widely used. It remains useful in a way that computer vision applications are not, due to the simple fact that with some basic instruction, anyone can bring the Hinman Collator to whirring life. Its function and controls are self-explanatory, in spite of the device’s initially off-putting appearance. Yet the collator has long since given up any claim to be cutting-edge. It is an ancestral machine, a technological memento mori. Even so, its invention could be said to mark the birth of the digital humanities, an honor usually accorded to the Index Thomisticus — a punch-card concordance of the works of the medieval theologian Thomas Aquinas begun in the 1940s. But where the Index represented Big Data avant la lettre, the Hinman Collator offered a kind of microscopic prosthesis. It was (and is) unique in being less a means of tabulating data than an instrument designed to augment the human eye. It made possible the detection of evidence that was otherwise difficult to discern. In this respect, it is more like Galileo’s telescope than, say, the relational database: it divulged new worlds.
Smith, Steven Escar. “‘The Eternal Verities Verified’: Charlton Hinman and the Roots of Mechanical Collation.” Studies in Bibliography 53 (2000): 131.
Hinman, Charlton.“Mechanized Collation at the Houghton Library.” Harvard Library Bulletin 9 (155): 132
Ibid.
Mongan’s report is in the Charlton Hinman Papers, Kenneth Spencer Research Library Archival Collections, University of Kansas, MS 69.
Hinman, Charlton. “Variant Readings in the First Folio of Shakespeare.” Shakespeare Quarterly 4 (3)(1953): 288.
Smith, Steven Escar. “‘Armadillos of Invention’: A Census of Mechanical Collators.” Studies in Bibliography 55 (2002): 139.
This essay is an edited excerpt from Rare Books & Ancestral Machines: A Handbook to the Posner Center for Special Collections at Carnegie Mellon University (Carnegie Mellon University Press, 2025).
