Among ciphertexts of the early modern period, this one stands alone for sheer visual weirdness. It may be found in the 1602 edition—and only that one specific edition—of a book about ciphers written in Latin by the illustrious “professor of secrets,” Giambattista della Porta. And it looks so utterly wacky that the typical reaction of modern commentators has been a kind of good-natured bemusement coupled with a conviction that, as an inscription, it’s surely not to be taken seriously. I mean, just look at it!
In The Ciphers of the Monks: A Forgotten Number-Notation of the Middle Ages (2001), David A. King characterizes this as “a code using ciphers randomly tilted to the vertical, as if the ciphers were coming home after a good party” (p. 232). I like King’s description, which conjures up whimsical images of the glyphs staggering drunkenly about, and building on it I’ve decided to refer to the system in question as “Della Porta’s Drunken Cipher.” As we’ll see, this name might be appropriate for other reasons besides.
The Ciphers of the Monks is a wonderfully interesting book which can, as of this writing, be downloaded from ResearchGate (otherwise it seems, alas, to be not only out of print, but wholly unavailable on the second-hand market as well—I’ve searched repeatedly to no avail). In it, King presents an exhaustive account of the Cistercian numerals, a late medieval and early modern system of numerical notation in which any integer from 1 to 9,999 can be written compactly as a single glyph. Each digit from 1-9 was associated with a distinctive form, which could then be given a place value by writing it in one of four quadrants around a stem (or bar, or stave, or whatever you want to call it), with one position representing units, another tens, another hundreds, and another thousands. The forms associated with the digits varied, as did both the orientation of the stem (sometimes vertical, sometimes horizontal) and the arrangement of the place values around it, but here’s the arrangement used by one online Cistercian Numerals Converter:
Note that the forms of the digits are mirrored across vertical and horizontal axes rather than just shifted “as is” into other quadrants. Although the Cistercian numerals were themselves designed to represent numbers, their distinctive strategy for combining digits into single composite glyphs was sometimes also harnessed as a means of enciphering combinations of letters of the alphabet, which is where Della Porta comes in.
Della Porta’s Drunken Cipher is an elaboration of a simpler alphabetic cipher that appears in all editions of his book De Furtivis [or Occultis] Literarum Notis, and not just the 1602 edition, and that bears a more obvious relationship to the Cistercian numerals. Of the simpler of these two ciphers, King writes (p. 232):
Although he [Della Porta] mentioned Agrippa’s ciphers [a version of the Cistercian numerals] as a means of representing numbers, the 22 ciphers which he presented (11 basic left-facing shapes repeated with a dot on the second set) are different from those of his predecessor…and his use of them in 30 ciphers representing 4-letter words, apparently without any linguistic significance, most curious. Alas, he gave no indication of what he intended with these combinations—compare Daniel Schwenter.
I’ll proceed to the simpler of Della Porta’s two Cistercianoid ciphers in a moment, but let’s first consider the Daniel Schwenter cipher which King mentions in the above passage (and describes on pp. 228-232). It appeared in a book entitled Steganologia et Steganographia published around 1620, which is to say somewhat later than Della Porta’s book. Like Della Porta, Schwenter mentions Agrippa’s take on the Cistercian numerals as a model, and he then proceeds to present a set of forms representing letters of the alphabet much as the forms of the Cistercian numerals represent digits, each confined to the upper right quadrant around a stem.
In the accompanying German text, Schwenter explicitly lays out his method of combining these forms into composite glyphs each representing up to four letters, much as the Cistercian numerals combine up to four digits. Specifically, he writes that he follows the sequence upper left, upper right, lower left, lower right, and indicates a word break by leaving a quadrant empty. Parsed in the way Schwenter describes, the sample ciphertext shown at the bottom above reads WER_ / GOTT / _VER / TRAU / T_HA / T_WO / L_GE / BAUT, i.e., “Wer Gott vertraut hat wol gebaut.” Clever, huh?
Notwithstanding King’s remarks, Della Porta actually describes his intentions in nearly as much detail as Schwenter does, albeit in Latin rather than German. Della Porta’s text was translated into English in the first half of the twentieth century, but the translation remains unpublished, so I’m forced to fall back here on my own humbler knowledge of Latin, with apologies for any infelicitous results. In the first 1563 Naples edition of De Furtivis Literarum Notis, at page 93, Della Porta writes:
Twenty or more characters may be found which, if they adhere to the top and bottom, left and right of an upright staff, thereby always retain their shape and power; and we draw first at the head of the staff which appears to our left, second to the right, third at the foot of the staff likewise to the left, and fourth to the right. We will now be poised to write four [letters], and we will show the matter more clearly by example. Here are the characters.
The same printing plate was used for the sample ciphertext in the expanded 1602 Naples edition, at page 133, a little worse for wear; but a new plate was substituted for the key. In the latter, the cipher letters no longer line up neatly with the plaintext letters above, which makes the key difficult to use. We also see the dot missing from the cipher letter for z and a potentially confusing break in the stem for the cipher letter for m, although canny readers could have inferred that these were unintended deviations from a straightforward pattern.Both plates were manually redrawn for other editions, starting with the 1591 London edition (also issued with a spurious 1563 imprint). Here’s its key:
And here’s the associated sample ciphertext:
This copy comes fairly close to the original 1563 plate, even unto the preservation of a break in the upper left quadrant of the twenty-second character; but there’s a dot missing from the upper left quadrant of the fourth character.
Della Porta’s work was also published three times under the alternative title De Occultis Literarum Notis (both titles translate equally to “On Secret Marks for Letters,” whatever divergent connotations furtivus and occultus might seem to have). The cipher we’ve been examining so far appeared in the 1593 Montbéliard edition at page 116; the 1603 Strasbourg edition at page 123; and the 1606 Strasbourg edition, likewise at page 123. In the key plate, the dot is missing from the cipher letter ordinarily assigned to o, making it indistinguishable from b. To complicate matters further, the plaintext alphabet is given incorrectly in the 1603 and 1606 editions, including k but not q, and so assigns the wrong ciphertext letters to plaintext letters k through p inclusive, so that the letter with the missing dot ends up associated with n. The illustration below is from the 1606 edition.
The layout of the key in the 1593 edition has the correct plaintext alphabet, but it presents an even worse coordination between cipher letters (broken into two lines) and plaintext letters (at the top in a single row) than the 1591 London edition, to the point that it’s scarcely usable at all.
Meanwhile, the sample ciphertext has been broken into three lines for these editions with a decorative flourish at the end.
This time there are quite a few discrepancies relative to the original 1563 plate in the presence or absence of dots. All the letters come in dotted and undotted pairs—A/N, B/O, C/P, D/Q, E/R, F/S, G/T, H/V, I/X, L/Y, M/Z—which means that a dot always makes a difference between one or another letter.
Nothing in Della Porta’s text explicitly states how the cipher forms in the key—all shown there as they would appear in the upper left quadrant—are supposed to be adjusted for the other three quadrants. From the sample ciphertext, it’s easy to see that letter forms are being flipped across the vertical axis when they appear in the upper right quadrant. Less obvious is what happens in the two lower quadrants—that is, whether the versions of the letters in the upper two quadrants are simply being shifted downward along the stem, or whether they’re being flipped across the horizontal axis to become mirror images of themselves. In most cases, the shapes of the letters found in the sample ciphertext could be interpreted either way, since flipping a letter vertically turns it into another valid letter—the pairs being B/C, E/F, G/H, I/L, and their dotted equivalents O/P, R/S, T/V, X/Y. Only M and its dotted counterpart Z aren’t vertical mirror images of other valid letters. And so by looking at how M appears in the lower two quadrants—there are no examples of Z in the sample ciphertext—we can tell that letters are being flipped across both the vertical and horizontal axes, like this:
Thus, we can infer that Della Porta is handling the transposition of forms into different quadrants in the same way as Schwenter, and following the lead of the Cistercian numerals. If we now use this scheme to read the letter forms in the sequence Della Porta specifies (upper left, upper right, lower left, lower right), we find that the sample ciphertext does contain a meaningful message, notwithstanding King’s characterization of it as “apparently without any linguistic significance.” In fact, the intended plaintext turns out to be the exact same one he had used to demonstrate a different cipher a few pages before: “Multis cladibus ultro citroque datis et acceptis universa pene civitas occupata est. Reliqua non scribam sed in congressum nostrum reservabo.” The four-letter plaintext blocks should thus be MVLT ISCL ADIB VSVL TROC ITRO QVED ATIS ETAC CEPT ISVN IVER SAPE NECI VITA SOCC VPAT AEST RELI QVAN ONSC RIBA MSED INCO NGRE SSVM NOST RVMR ESER VABO. Here, for comparison, is what actually got printed:
1563: MVLT ISCL ADIB VSVL TEOC ITRO QVED ATIS ETAC CEPT XSVN IHEE FAPE AECI VITA SOCC VPAT AEST RELI QVAN ONSC RIBA MSED INCO NGRE SSVM NBST EHMR ESER HABO
(=”Multis cladibus ulteo citroque datis et acceptxs uniheefa peae civitas occupata est. Reliqua non scribam sed in congressum nbstehm reserhabo.”)
1591: MVLT ISCL ADIB HSVL TEOC ITRO QVED ATIS ETAC CEPT XSVN IHEE FAPE AECI VITA SOCC VPAT AEST RELI QVAN ONSC RIBA MSED INCO NGRE SSVM NBST EHMR ESER HABO
(=”Multis cladibhs ulteo citroque datis et acceptxs uniheefa peae civitas occupata est. Reliqua non scribam sed in congressum nbstehm reserhabo.”)
1593 (using 1593 key): MVLT ISCL ADIB VSVL TROC ITRO QVED ATIS ETAC CEPT ISVN IVRR SAPE NECI HITA SOCC HPAT AEST RELI QVAN ONSC RIBA MSED IACO AGRE SSVM NOST RVMR ESER VABO
(=”Multis cladibus ultro citroque datis et acceptis univrrsa pene cihitas occhpata est. Reliqua non scribam sed ia coagressum nostrum reservabo.”)
1593 (using incorrect 1603/6 key, just to see what happens): LVKT ISCK ADIB VSVK TRNC ITRN PVED ATIS ETAC CEOT ISVM IVRR SAOE MECI HITA SNCC HOAT AEST REKI QVAM NMSC RIBA LSED IACN AGRE SSVL MNST RVLR ESER VABN
(=”Luktis ckadibus uktrn citrnpue datis et acceotis umivrrsa oeme cihitas ncchoata est. Rekiqua mnm scribal sed ia cnagressul mnstrul reservabn.”)
The 1591 plate contains the same ten errors as the 1563 plate and introduces one new one (cladibus becomes cladibhs). By contrast, the 1593 plate contains just five errors, all different from the errors in the 1563 plate, so it would seem that the 1593 plate wasn’t copied from the plate in the 1563 edition and was instead prepared independently from a manuscript source. The only other attempt I’ve seen to make sense of this sample ciphertext appears on S. Tomokiyo’s Cryptiana site. It makes it through the first three and a half cipher characters of the 1591 plate, inserting plaintext letters into the image as “Multis cladibus” (even though the u is actually written as an h), and then stopping, maybe put off by the mistake.
As I’ve already mentioned, the simpler of Della Porta’s two Cistercianoid ciphers appears in all editions of his book, but the more complicated, “drunken” cipher is found only in the 1602 edition. S. Tomokiyo’s Cryptiana site refers to the latter as a “further imaginative enciphering” and reproduces it in facsimile, but with no attempt at decipherment. The “drunken” cipher differs from the simpler one in that it introduces two additional keys, both of which lack not only k, like the other key, but also y and z. For clarity, I’ll refer to the key we’ve been discussing so far as the first key, and its forms as first-key letters. Della Porta states nothing explicitly about how the two additional keys are to be implemented, writing only that the idea is now to make each character represent six letters rather than four as previously, and then leaving the examples to speak for themselves as best they can.
The second key consists of ornaments (my term) incorporated into the stem of the character:
Various irregularities can be spotted here, but the least clearly drawn letter of the lot is r, which looks as though it might consist of three quarters of a circle in the middle of the stem, as though someone had tried to jump through a hoop that was a little too small and ended up ripping it apart and leaving a piece dangling l suspect it’s actually supposed to contain a complete circle, analogous to the ones seen in a and b, since a half-circle would make it identical to u.
Some of the rotations are forty-five degrees or less, and in these cases it’s also easy to infer a direction: for example, counterclockwise for b and d and clockwise for c and e. However, we can’t tell from the key whether the horizontally oriented letters f and q through x involve rotating the character clockwise or counterclockwise, which would make a difference for the reading of all the other letters in a given character. Meanwhile, an even more serious source of ambiguity is that the feet added to the tops and/or bottoms of stems for letters g through x are liable to be confused with parts of certain letters in the first key. If a first-key a, c, or e were to fall in the same quadrant as a foot, it would be impossible to distinguish the combination from a first-key e, d, or q respectively. Moreover, it’s unclear what would happen if a foot were to fall in the same quadrant with a first-key b, d, e, o, q, or r, since both would require a line in the same exact place.
The bottom line is that Della Porta (presuming he was responsible for this addition to the text) doesn’t seem to have thought the extended scheme through very well. But that sample ciphertext is so wonderfully odd that I just can’t look away. Can we unriddle its secrets and expose what it’s trying to do, and how?
Let’s start with the hypothesis that Della Porta used the same plaintext here that he’d used for his other, simpler ciphertext using just the first key, and which he’d already carried over without comment from his discussion of yet another cipher: “Multis cladibus ultro citroque datis et acceptis universa pene civitas occupata est. Reliqua non scribam sed in congressum nostrum reservabo.” This time, he would have had to break that message up into six-letter groupings rather than four-letter groupings. If he’d done that, he should have ended up with twenty six-letter groupings: MVLTIS CLADIB VSVLTR OCITRO QVEDAT ISETAC CEPTIS VNIVER SAPENE CIVITA SOCCVP ATAEST RELIQV ANONSC RIBAMS EDINCO NGRESS VMNOST RVMRES ERVABO. A problem immediately arises: the sample ciphertext contains only nineteen characters, not twenty. But let’s not give up too hastily.
I started my analysis with the fourth character from the left in the top row:
I decided that the diamond in the middle of the stem could only be q, according to the second key, implying that the next letter would be v. The horizontal orientation of the stem is consistent with v in the third key, but that key also shows two feet extending downwards from the stem for v, not upwards as in the ciphertext character. This led me to suspect that the character might have been printed upside-down, and that it should actually be:
I reasoned further that if the whole of the top row of the sample ciphertext had been similarly flipped upside-down, then this character would actually be the fifth from the left, and not the fourth. The fifth expected six-letter grouping would be QVEDAT, which contains q and v. And if the horizontal orientation of the third-key letter v had been achieved by rotating the character counterclockwise, the four first-key letters could be read as EDAT, running from right to left and top to bottom. This order is different from what we saw in the earlier, simpler ciphertext, in which letters were read from left to right.
The d is also problematic because one of its lines is simultaneously one of the feet belonging to the third-key letter v. Without that line, it would be a c, while factoring it out of the third-key v would turn that letter into q. There’s no way to rule out the readings QQEDAT and QVECAT.
But even with all those caveats, this is still a plausible interpretation of what the character is meant to be. It also makes some logical sense that the stem would be drawn first, and that it would therefore encipher the first two letters of the six, with the remaining four letters added around it afterwards.
Now let’s see whether this same approach fits the rest of the message. Working with the top row of the sample ciphertext flipped upside-down, the first character from what has now become the left should be MVLTIS. The second letter, v, is the same as in QVEDAT, so the third-key component of the character should be the same: a counterclockwise rotation plus two downwards feet. Here’s my reading of what we see:
There are a few surprises here. First, the m doesn’t take the form we’d expect from the key. The ornament itself is correct, zigging and zagging in the right directions, but the third key shows it placed at the bottom of the stem for m rather than the top. At the top of a stem, the same ornament is instead supposed to represent p. But this discrepancy between the second key and the ciphertext turns out to be consistent: whenever the second key shows an ornament at the bottom of a step, the sample ciphertext always places it at the top, and vice versa. I infer that the second key itself must be wrong, and that it should have been printed like this:
Another surprise is that the left foot of the third-key letter v ends up appearing close to the horizontal center of the character. It seems that whenever a second-key ornament is attached to an extremity of the stem, the stem is understood to end wherever the ornament joins it for purposes of adding additional marks, whether these are parts of first-key letters or third-key feet.
The third surprise is that the final four letters, LTIS, read from left to right as in the earlier, four-letter-character ciphertext, and not right to left as in QVEDAT. The right-to-left reading in QVEDAT turns out to have been an anomaly (although we’ll see it again) and was presumably a mistake. The usual rules are evidently:
- Encipher the first letter using the second key (flipped upside down from the way it was printed).
- Encipher the second letter using the third key, and if the stem needs to be horizontal, rotate it counterclockwise.
- Encipher the remaining letters using the first key, starting with the third letter at upper left.
- Encipher the fourth letter at upper right.
- Encipher the fifth letter at lower left.
- Encipher the sixth letter at lower right.
The second character should read CLADIB. It almost does, except that the d is wrong; as written, the character would actually read CLAFIB. The first letter, c, is again written upside-down relative to what’s shown in the second key (as we’ll continue to see), and the final four letters again read from left to right. The l and the b both require a mark in the same place, as did the v and d in QVEDAT, but this time we find two marks there rather than just one. In the diagrams that follow, whenever a letter has been written using an incorrect form, I’ll transcribe it in green, and I’ll also use green to show how I believe the ciphered letter should have looked, overlaid on the red marks that highlight what was actually written.
The third character should read VSVLTR. We run into another mistake with a first-key letter that, in this case, is upside-down—as written, the character would read VSTLTR—but there are no surprises otherwise.
The seventh character should read CEPTIS. As written, it actually reads CESTIE, due to an extra stroke on the p and the s being written upside-down without its dot (or at least that’s the simplest way I can find of explaining what went wrong). The third-key e rotates the character around thirty-five degrees instead of forty-five, as for c, which is an awfully subtle difference.
The eighth and final character in the line should read VNIVER, but the r is missing its dot.
The bottom row of characters in the sample ciphertext doesn’t appear to be upside-down like the top row. I suppose it’s possible that the change in orientation between lines was intentional—a kind of quirky variant on boustrophedon writing, like this:
But considering the many other more definite mistakes in the sample ciphertext, I suspect this was just one more of them. The first character in the new row should read SAPENE, but we can read it that way only by violating several of the rules laid out above.
The first letter has been enciphered using the third key rather than the second as usual, and it’s rotated into a horizontal orientation clockwise rather than counterclockwise. The second letter has also been enciphered using the second key, so the protocols for enciphering the first two letters have been switched around. Finally, the four first-key letters are read right-to-left rather than left-to-right. Maybe the cipher wasn’t the only thing that was drunken here.
The next character should read CIVITA, but the foot on the third-key i has been drawn in the wrong direction, towards the right rather than towards the left. Because this mark overlaps the first-key letter a, it ends up garbling that letter as well—although I suppose the a is arguably just a casualty of the general ambiguity over third-key feet, and not a mistake in and of itself.
Next should come SOCCVP, but the four first-key letters have been written from top to bottom and then from left to right, rather than the other way around, and the p has also been written upside-down.
Next should come ATAEST, but the foot is missing from the third-key t, and the four first-key letters run from right to left.
Next should come RELIQV, but a dot is missing from the q. This character confirms the way I’d speculated a second-key r was supposed to be drawn, as a circle in the middle of the stem—you may recall that this letter wasn’t written very clearly in the key itself.
Next should come ANONSC, but the s is upside-down and the horizontal line at the top hasn’t been rotated along with the rest of the character.
Next up is RIBAMS. The first letter isn’t the same as the second-key r seen in RELIQV, nor does it have anything in common with the confusingly-drawn r in the key itself; instead, it’s written as a second-key e. Two of the first-key letters are also written incorrectly, with an extra stroke in the a (turning it into f) and an m that doesn’t loop back all the way to the stem (thereby turning it into l). As written, the character would read EIBFLS.
Next is EDINCO, but the o is missing a dot. (Had Della Porta possibly confused the beginning of this sequence with the beginning of the previous one, and might that be why he enciphered an E rather than an R?)
And then NGRESS, but the e has a spurious dot.
Next comes VMNOST, missing dots from o and t.
Finally, we come to RVMRES. The first letter isn’t the same as the second-key r in RELIQV, but it does resemble the confusingly drawn r in the key itself. The first-key r is missing a dot.
The twentieth six-letter grouping, ERVABO, is missing from the printed ciphertext.
Presented in isolation, Della Porta’s sample “drunken” ciphertext might have been uncrackable, not because of any particular ingenuity in its construction, but simply because it’s so riddled with mistakes. Some of the mistakes may have been introduced by whoever prepared the printing plates, but others seem as though they must have been present in the manuscript source. But enough of the ciphertext behaves as expected to show that it’s not meaningless gibberish, and with a little more care and practice, I’m sure this could have functioned as a viable (if clunky) cipher. Here’s my own attempt to recreate Della Porta’s sample ciphertext with mistakes corrected and inconsistencies or ambiguities reduced (for example, I’ve tried to draw all “feet” half-length or shorter). I’d be surprised if I haven’t made a mistake of my own somewhere, though.
One improvement might be to substitute curved stems (bowed leftwards or rightwards, or upwards or downwards) for the letters beyond F in the third key, so as not to interfere with the first key as the “feet” do; and maybe to introduce a zig-zag stem or two parallel stems as well, so as to avoid subtle distinctions of slope that differentiate B from D, C from E, and so on.
The rotation of the stem is arguably the most distinctive feature of the Drunken Cipher, and the one that gives it its “drunken” appearance; but in resorting to rotation, Della Porta wasn’t entirely alone. Girolamo Cardano, writing a few decades before Della Porta, had suggested extending the scope of the Cistercian numeral system beyond 9,999 by—among other things—assigning meanings to figures rotated into a horizontal position as well as into clockwise and counterclockwise diagonal positions (see King, p. 210). Moreover, the rotation of stems also forms an important part of the Characterie of Timothie Bright (1588), sometimes identified as the “father” of English shorthand. A reprint of Bright’s work may be found here, although it seems to present some of the characters incorrectly (in the original 1588 publication, these were drawn in by hand rather than printed).
What I’ve been calling the stem, Bright calls the “bodie,” which has four quadrants referred to in terms of the left and right “head” and the left and right “foot.” Bright’s system differs from Della Porta’s in that the left and right sides are used to differentiate between letters rather than being used to encipher two different letters simultaneously (much like earlier versions of the Cistercian numerals, which had used both sides at the top of the stem for units and both sides of the bottom at the stem for tens). Here’s Bright’s basic alphabet (the next few illustrations are borrowed from a summary presented by Edward Pocknell in 1883):
Six more forms can appear in either of the lower two quadrants, to either side, but these have no similarly fixed meanings and are instead assigned to a constrained vocabulary of words beginning with whatever letter is indicated at the top of the stem, ordered alphabetically. Here’s the list for A:
Rotation then comes into play whenever more than twelve words beginning with the same letter need to be accommodated: 13-24 ninety degrees counterclockwise; 25-36 forty-five degrees clockwise; 37-48 forty-five degrees counterclockwise. Here, by way of illustration, are words 13-24 in the A category:
Like Della Porta’s Drunken Cipher, then, Bright’s system makes use of rotation to expand the number of visually distinctive states that can be assigned distinctive meanings, albeit only by a factor of four, so that they don’t comprise a whole new alphabetic key of their own. Thus, each of the glyphs described so far simultaneously indicates a letter of the alphabet, a duodecimal number, and one of four rotational states corresponding to the next-higher numerical place.
Bright’s method of writing numbers, so far as he explains it, doesn’t show any obvious influence from the Cistercian numerals.
NOmbers are writen by the heades of the compounde Characters, with a ſtreight bodie hanging, and take increaſe by place, as Ciphers in Arithmetike:
There’s a space after the colon for examples to have been written in, but the Bodleian library copy doesn’t show any, and the Ford edition muddies the waters by ending the sentence with a period. Still, Bright seems here to be describing the use of an “ordinary” place system. Based on that observation, we might doubt whether his system as a whole bears any meaningful resemblance to the Cistercian numerals at all. But he goes on:
Propper names, if they be ſignificant, are written by character: as, fielde, day, &c Otherwiſe the head of the character bearing the figure of a letter added alſo to the foote, and ſo ioyned in one figure, may ſerue for two letters: as,&c. And ſo other two, till all the worde, or as much as is neceſſarie, ſhal be written: with a marke at the ſide of the firſt Character, to ſhewe that it is a name.
Thus, with proper names written phonetically, a second letter could be added at the bottom of the stem, rotated 180° from its usual form when appearing at the top. This provision approaches much closer to the logic of the Cistercian numerals, even if it isn’t a core feature of Bright’s system. Still, as far as I can see, there shouldn’t be any situation in which Bright would combine this method with stem rotation. So Bright uses a few of the same building blocks as Della Porta’s Drunken Cipher, but he doesn’t deploy them in anything like the same way.
I hope you’ve enjoyed this little exposition of a fun and neglected historical cipher. Please don’t send me any long messages written in it, though—when it comes to reading, a little Drunken Cipher goes a long way!