SEIZING ORDERED CHAOS
A Serious Attempt at Reproducing Gutenberg's B-42 Types
by Alan C. Waring
IIn celebration of the "Man of the Millennium," a Tokyo publishing house nobly decided to establish a printing museum, which opened in October, 2000. Over two years before the museum's planned opening, the Tokyo firm actively started its acquisition of wooden handpresses and suitable types. They commissioned two different Americans to build wooden presses, a reproduction of the Plantin-Moretus press, and a press like those thought to be used in Gutenberg's workshop. They were successfully built and delivered. It was at this point that problems began for the Tokyo firm, as their plans included purchasing a reproduction font of Gutenberg's B-42 types as seen in his monumental 42-line Biblia Latina. They could have settled on any number of readily available blackletter fonts, but to their great honour, they decided differently. It is ironic that a Far Eastern publishing house thought so highly of the Western invention of leaden movable types (and the attendant wooden handpress), that they deemed their printing museum incomplete without a reproduction of this historic font.
Their search began logically in Germany. Surprisingly, no one could cast this font, as only an incomplete font of lead matrices exist. The search then went to England where the story was much the same, no extant matrices and no working foundry.
The Western emissary tasked with the acquisition of type & presses began to feel the squeeze of time. With only seven months left to acquire the types and to set four pages of the Gutenberg Bible, he spoke to Mr. Stan Nelson of the Smithsonian Institution's Graphic Arts Division at the National Museum of American History. Mr. Nelson suggested that he contact Mr. Theo Rehak at the Dale Guild Type Foundry in Howell, New Jersey. By the time negotiations were finalized, the delivery date was just five very short months away! Eighteen months would have been more realistic, considering the fact that a workforce of two would be doing the job. Briefly, this entails procuring the exemplars, producing the artwork, matrix production, casting, milling types to Gutenberg's body-size and packing for overseas shipment.
After surviving the initial shock of Mr. Rehak's agreement to a ludicrous deadline, I realized that we had never done a job of this size, nor reproduced a font under the ax of a deadline. In order to avoid failure, cardiac arrest or both, I made a series of production schedules which would, at the very least, enable our minds to concentrate on a manageable segment of the whole. But before this could be done, I had to discover just how many of the approximately 270 B-42 characters would be needed to set the four pages chosen by the Tokyo firm. Fortunately, we needed only to produce approximately 160 characters. Had the entire suite of characters been needed, the deadline could not have been agreed to, and the composed pages would not have been ready for the October opening.
Gottfried Zedler's character-set, published in Mainz, (1929) proved an invaluable aid in drawing up the production schedules. I divided the job into six character groups and worked through the four pages, following the sequence of characters as they appeared line by line. In this way, types could be shipped as soon as we had finished the characters for the first page. As it turned out, this provided the additional time needed to compose the four pages in a timely manner. To insure we did not overlook any characters, schedules were prepared for the following: artwork completed, master engraving patterns, matrix cutting, 1st casting of types, alignment & set-width, number of sorts per character, milling, and types shipped. It should be noted that there were approximately ten characters that occurred only once or twice upon the four pages to be set, but the only measurable savings in labour with these characters was realized in milling the types to the original body size.
It might be as well to candidly state the following before I go any further into this narrative; the typeface that Theo and I have seized in metal is close to Gutenberg's B-42 types, but they are not his types. Without the original punches or matrices, no one can cast the types seen in the 42-line Bible. Even if we had access to the those punches and matrices, we could not cast all the B-42 types, as it must be assumed that punches failed while matrices were driven, thus necessitating the recutting of replacements. These replacements would be close to their predecessors, but would and do show discernible variations. There is strong textual evidence for this. If only the matrices survived, it must be asked, what matrices? While the evidence shows that most of the original punches survived the printing of all 1,282 pages, very few original matrices would have been usable past the setting of the first volume if they were copper, and if, as some scholars believe, they were leaden, almost none. If unused original types miraculously survived, these artifacts must be considered third generation to the punches and matrices. And again it must be asked, what types? Taking into account the foregoing inevitable generational instability of the typefounding process, even Gutenberg could not reproduce Gutenberg character-for-character.
There is also evidence that a few characters were recut early in the printing process as they may have displeased Gutenberg's finely honed aesthetic. This has been the experience of many type designers, but it is usually done before the use of the types. When want of time and large sums of money are involved, not always. This has been our own galling experience. Our B-42 Blackletter exists in two states, and these exist in artwork, matrices, type and printed specimens. This is to say that we make no other claim than that we have made a whole-hearted attempt at reproducing the B-42 types, and we think that our font could not be confused with any other antecedent.
Preparing the artwork for the cutting & casting of a typeface is never an easy endeavour, but reproducing a typeface from extant types or punches as working exemplars is far more easily accomplished than working from the printed page or the freehand drawing of a completely new font. While the former is a task requiring 300-400 hours of uninterrupted labour, it is a pleasant task which affords the leisure to contemplate the mind and the skills of the designer while drawing. The latter two incite mixed emotions; the happy anticipation of the challenge along with the weight of having to tread a long and grueling road on one's own.
To embrace a job of this magnitude in an age when the number of letterpress printers shrinks by the minute, a romance for the craft must be deeply felt. While we may be poor businessmen, we share an admiration of those who have gone before us. We also share an appreciation for the simplicity of late mediaeval life and we enjoy the freedom to engage in a challenging and personally rewarding craft. We actually get to see the fruits of our labours, and then use them as we choose. Our reality is that there are very few who appreciate or require our particular anachronistic skills.
Years ago, I engaged in a flight of fancy about reproducing Jan Van Krimpen's tour de force Cancelleresca Bastarda. I became aware of a seemingly endless number of variant characters in that fount. Since then, I have provided the artwork for seven or eight typefaces which contain the average number of characters. I now realize that the neural strain of such work takes it toll and tempers one's ambitions. As fond as I am of blackletter types, I had never considered the B-42 font for reasons too numerous to mention here, but I had to ask myself if I was prepared for such a monumental task? The average fount of foundry type consists of about eighty characters. I was confronted by three times that figure, and then some (my math seems errant, but knowing Theo as well as I do, I knew he would want to finish the font after the Tokyo contract was completed. My assumption was correct). The and then some also means, in terms of labour, many, many more, by reason of the numerous two and three letter ligatures found in the B-42 font. Oft times, these have contraction marks, further adding to the time it takes to complete a single drawing.
Unbeknownst to Theo, I began the artwork the day he called with his first mention of the project. Said he, "Don't worry, it won't happen anyway." Yet I knew that we could never finish the project on time if I waited until we received the go-ahead. The word came six weeks later, by which time we had cut and cast eight trial characters for the Tokyo firm to examine. Before the contracts were signed, I was well into the artwork. If the project fell through, I would still have the benefit of an enjoyable study in blackletter types. These eight trial characters were ultimately rejected and redrawn by the "Art Department" upon the arrival of superior exemplars, yet much was learned in their production.
We are not in the enviable position of having in our keeping a copy of the B-42 Biblia Latina, nor do we have free and continual access to the same, a very necessary thing as I will explain later. While not without original specimens, excellently produced second-generation materials are relatively easy (though not inexpensive) to obtain. I have utilized photos, scans, a few revealing half-tones and a complete facsimile reproduction. By far, the most useful has proved to be the 2 volume Idion/Verlag facsimile collotype reproduction, (Munich 1977-78). This is the finest facsimile ever produced, using one of the best extant copies of the Gutenberg Bible. No expense was spared in its production. The quality of my work is largely based on the quality of this facsimile.
When preparing artwork from very early printing, the lettering artist has to contend with many variables not found in any other situation. These variables dictate the necessity of having as many examples of the individual characters as is possible.
Once sufficient exemplars are in hand, it is a relatively simple process to isolate the individual images from their surrounding fellows. The real difficulties begin when I am forced into making a choice between variant images of the same character, real or apparent. Subjectivity is never a good thing when attempting to reproduce an early typeface, especially when the exemplars so often exhibit muddy impressions. Under-inked or over-inked characters, worn or poorly cast sorts of the same character all appear on the printed page differently one to another. The operative word here is appear. The eye in co-operation with one's judgment must come to terms with having to make this choice. The choices I had to make were much the same as those Gutenberg and his punchcutter/s had to make from their manuscript exemplar 550 years ago. That did not make it any easier. I agonized over many of them. After the contract was fulfilled, I redrew several characters, as they did not please me upon further examination in trial settings. My eye had become more acute with the time spent drawing character after character. To Theo's credit, he never chided me for the extra work and expense I had put him to. After the first sixty to seventy drawings I finally felt some degree of comfort with the font, but I was never able to relax, though I began to find myself engaged in long conversations with Meister Gutenberg about six weeks into the project.
The format to which I work is 7.5 inches in height. That is, my finished drawings are 7.5 inches from top of ascender to bottom of descender. For this typeface, the proportional relationships of the capitals, ascenders and descenders to the x-height had already been established by Gutenberg or perhaps by his manuscript exemplar. The size of format insures crisp and proportional reductions for the production of the master engraving patterns and finally the types.
The next step caused me great difficulties. Regardless of the degree of sharpness which my choice of images showed, I had to deal with the ink-squeeze. When a typical character is enlarged to my working format (see Blackletter page), it looks as if the ink were roofing tar applied with a trowel. Many of the characters appear so seldomly in the text that the choices I had for exemplars were less than helpful. In general, I used five or six images of a given character to produce a finished drawing, using the best or cleanest elements of each, thus establishing the parameters of the metal type beneath the ink, stem by stem and serif by serif. More than once I threw up my hands in extreme frustration. Many so-called finished drawings were rejected because I would later find a truer image of a given character. It took about 200 drawings to secure the 160 needed to fulfill the Tokyo contract. In a discipline which demands objectivity, this step in the overall reproduction process could not be accomplished without some degree of subjectivity. Ultimately, I had to decide where the ink-squeeze stopped and the edges of the type began. This is an impossible task without the practical experience of letterpress printing coupled with a constant scrutiny of ones own work. I always strive for proper inking with loupe in hand. By proper I mean solid black images short of muddied edges. I have tested many inks with the same forme looking for that which produced the ancient velvety black that sets apart early printed books from most others. Mixing my own inks has produced varied results. Tweaking inks with pigments, binders or extenders has given the best results. These tests were done on both dry and dampened papers, as well as vellum. Most of my printing is done on dampened papers as they give the best results with the least amount of ink. Less ink equals less squeeze which yields sharper images. After the first few hundred hours work, knowing when to stop the removal of ink squeeze became somewhat easier.
Most of the foundry type I use is new and as a part-time type designer I cannot resist examining types under a loupe, especially those borne of hand-cut punches. I also enjoy taking existing types (sets of initials, or complete founts, sort by sort) and opening them up with hand-gravers, producing inline or outline variants. I have also had the privilege of examining suites of hand-cut punches which show various degrees of ability on the part of the punchcutter. Drawing upon my work with Theo in resurrecting several lost type designs over the last ten years, I had to summon up all my experience in our attempt to do justice to Gutenberg and his punchcutter/s.
Along with the experiential, I tried to keep in mind Gutenberg's quest for perfection and the exemplar from which he worked, tempered by the finished characteristics of punches cut in the mid-fifteenth century and types cast in a handmold. His Bible amply testifies to his quest. Its overall beauty cannot be denied. The vast improvement of the B-42 types is seen in the appearance of kerned characters and the multitude of new ligatures which were not found in Gutenberg's two earlier Strassbourg fonts. Each one of these new ligatures also required the cutting and casting of an alternate character. "The narrower letters were used to contract the length of a line and the wider ones to expand it. The space separating words remained constant." And again, "…the narrower or wider alternate characters were skillfully deployed to yield lines of identical optical width…" (Kapr *) These are faulty conclusions. Either the wider or the narrower characters are alternates. As the textual use of the wider characters out-number the narrower ones by approximately 3 to 1, the latter must be considered the alternates. What usually makes these alternates narrower is the fact that they lack serifs on the left side. They are never used or omitted to shorten or lengthen the space required by a given number of characters within a line. Gutenberg had employed them in both his earlier founts, the Sibyllenbuch and the Donatus/Kalender, but none of the extant printed specimens used justified lines. Their inception and use is purely aesthetic. They do not represent consciously narrowed letterforms. These alternates were cut and cast to create tighter fitting of certain character combinations that produced too much white-space within words. The alternates (left side sans-serifs) always follow C, E, F, X, Y, c, e, f, g, r, t, x, and y. Nor is their use directly related to an attempt to follow the manuscript exemplar, but because of the inherent limitations of type when compared to calligraphy. The scribe can easily write over a portion of the previous letter to adjust the visual aspects of letters within a word. With leaden types, the compositor must take up his file, or the head of production can have another suite of punches cut, matrices driven & justified, and sorts cast. The beauty of well-written manuscripts goaded Gutenberg into cutting these additional characters, having decided his Bible would rival that of the scribe. Compositionally, justified lines made their first appearance, coupled with hanging punctuation. These technical advancements produced blocks of text optically superior to his exemplar, not a task easily accomplished! While the printed specimens prior to B-42 show equal word spacing, such is not the case in the 42-line Bible. It ranges from 2.75 to 5 Anglo-American points. If word-spacing exceeded five points, the words in that line were deftly letter-spaced. Of interest is that fact that the alternates are used regardless of the word spacing used in a line. If lack of space became a problem, contractions were used. Additionally, his decision to reset the 40 line pages to 42 lines (known as the Bible's first state and second state) is, in my estimation, the strongest evidence of this quest. Moreover, I believe that his decision to extend the pages to 42 lines after folio 5v in the Bible's first state was not done for economy. "Why did the compositors apparently change their minds in midstream and increase the number of lines to the page? The decisive factor can only have been an economy of about five per cent in the quantities of vellum and paper required." (Kapr **) If one examines the overall appearance of the forty line pages to the forty-two line pages, and counts the number of contractions used per page, one will discover 22.5 percent fewer contractions on the latter pages, thus yielding an increase in aesthetic value, not to mention text that is more easily read. This would seem to indicate that Gutenberg was attempting to follow a manuscript exemplar page by page, and that his motive was to improve readability as he moved from forty to forty-two lines.
Regarding his exemplar, it is more than likely that he chose a high quality manuscript written by an accomplished scribe. The text would have been consistently beautiful and sharply written.
As to the punches, they would have been less consistent and not as finely finished as those cut even five years later. I have no doubt that the punchcutter/s made every effort to retain the uniformity of the exemplar, and this work was done admirable well. But there is in this early typeface, greater variation in weight, proportion and x-height than would be tolerated even a few years later. "The Bar" was often raised by punchcutters right into the mid-twentieth century. Still, these variations lend an organic charm that in mass, produces an undeniably vigorous page.
Further, I had to take into consideration that the images on the printed page were produced by types cast in a handmold. Types cast thus are not always faced-up completely, and considering tens of thousands of types were cast, these anomolies will be seen in the printed images. Some of these may be flawed to the point that they will produce what appear to be variant characters. This fact can reach epidemic proportions swelling the numbers of supposedly unique characters dramatically. Even modern automatic casters can and do produce flawed types. Well-cast types from a handmold are not easily obtained, then or now. Lacking a mechanically pressured delivery of the alloy, types cast in a handmold are more likely to have incomplete faces, even if marginally so. The castor will not produce solid unflawed types until the mold & matrix reach their optimal casting temperature. These flaws show themselves in the printed images and have to be eliminated from the type gene-pool. This sifting through the great mass of available images proved to be the most difficult part of the artwork process.
Another unrelated difficulty was the multitude of filed characters. Having done more than my share of filing types in order to expedite the fitting of a new typeface, they were fairly easy to recognize, although some were skillfully done. A few characters shown in the Zedler list appear as filed sorts.
Of equal importance is the additional restraint which I imposed upon myself; that of being ever mindful of the lack of mechanical perfection found in the B-42 letterforms. Weight, proportion and x-height are but relative terms when dealing with this font. The gods (read demons) of uniformity beset me within and without. It required constant discipline to approach each character on its own merits. Otherwise it is exceedingly easy to translate the characteristics of one letter to another. This is unusually important, as there are (not counting ligatures) 8 c's, 10 e's, 10 i's, 11 q's, 12 t's & etc. When we first approached this font, we naively imagined that we could use one drawing and one matrix to produce both the full-set and its alternate fellow. We soon discovered the contrary. Individual punches were cut (perhaps by more than one workman) for each of these, and thus show quite a variation in the dominant letterform.
Having roughed out the finished forms, I then use a dip pen and sumi ink to define the finished parameters which will be translated to the types. With a properly prepared nib and an agreeable paper, a very sharp line can be drawn. Lacking unerring perfection with said instrument, I clean up the edges where needed with Chinese white and a red sable brush.
The 7.5 inch format is now reduced to 4.75 inches. These images are then trimmed tightly side to side, and put together in groups that total no more than 9 inches in length. The characters are then taped together side by side while set in a jig made for this purpose. These montages are sent to our photo engraver who then produces the master engraving patterns for the Guild's Benton Engraving Machine.
For this project, the montage work was not as straightforward as most. The trimming of the images is done on a lightbox with a fine grid. Normally, each fount has its own relationship to the vertical axis, which once discovered, may be adhered to. Such is not the case with the B-42 fount. Each letterform had to be visually checked with the exemplars, as the originals do not always conform to a uniform orientation.
The same holds true for the floating baseline, as we came to call it. Preliminary alignment is inherent in the montage work, because the placement of the images is translated to the engraving masters which is then translated to the matrices. Once again I checked the exemplars. If a character showed itself consistently high or low in Gutenberg's composition, that is where I placed it. While adjustments may have to be made in the matrix once types are cast, attention to detail in the montage work shortens our fitting work in the long run.
FINAL ALIGNMENT & FITTING
Alignment of the characters was treated above. To wit: the letterforms were aligned to the floating baseline as found in the exemplars. The fitting also presented us with unusual situations. Not infrequently we were faced with letter-spaced lower-case, and sorts filed by compositors as they set their lines. The former was done with subtlety, but it was easy enough to distinguish, as repetitive words show up at various lengths. The latter presented difficulties not easily dealt with, so four new characters, which occur most often in the text, were drawn, cut and cast to aid the compositor, thus eliminating much of this compositional work.
All in all, if I had known beforehand what numerous devils lay in wait during the preparation of the artwork for this fount, I would have disconnected my phone after Theo's first call, and moved to some obscure mountain top. Levity aside, I count it a great privilege to have worked on this project, and to do so with such a consummate craftsman as is Theo, my friend and fellow in a living craft. Setting these types sort for sort and line for line was and continues to be a sublime experience that never fails to transport me back to the Mainz Workshop, which has for so long remained shrouded in mystery and in many respects, still does.
Kapr, Albert, Johann Gutenberg, The Man and His Work. Translated from the German by Douglas Martin, Scholar Press, Hants, 1996
* page 146