On Thu, May 16, 2019 at 5:49 PM Jörg Rhiemeier <[log in to unmask]>

> On 16/05/19 20:45, Mike S. wrote:
> > On Wed, May 15, 2019 at 3:35 PM Jörg Rhiemeier <[log in to unmask]>
> > wrote:
> >
> >> I am puzzled
> >> how such a thing could pass the peer review board of a journal dedicated
> >> to Romance language studies!
> > One of the things that entered my mind when reading the abstract a few
> days
> > ago was how or why a competent peer review board would countenance
> > barbarisms like "quintiphthong" and "quadriphthong/quadraphthong" (it's
> > spelled both ways in the paper!), never mind the non-standard meaning
> > assigned to -phthong.  Admittedly, I am not an academic and do not know
> how
> > the peer review process works.
> I don't know, either. But a journal dedicated to the study of Romance
> languages should have a peer review board where at least a few members
> know what "Vulgar Latin" and "Proto-Romance" are, and are thus able to
> tell that a claim to a 15th-century text being written in that language
> must be spurious. This is a serious blemish on the reputation of the
> journal!

Indeed, and it's a bit baffling to me that this happened!  By the way, I
have a little update on the paper.  Apparently, the University of Bristol
is acting to distance itself from it.  They have taken down all posts on it
in their news feed (there was at least one last year and one this year) and
have put up the following message in their place:

16 May 2019

Please see below an updated statement.

Yesterday the University of Bristol published a story about research on the
Voynich manuscript by an honorary research associate. This research was
entirely the author's own work and is not affiliated with the University of
Bristol, the Faculty of Arts or the Centre for Medieval Studies.

The paper was published in the journal 'Romance Studies' following a double
blind peer review process in line with journal policy
a process used to validate the research quality of a study.

When a member of our academic community has a paper published in a
peer-reviewed journal, the University’s Media Team will determine whether
the findings are of public interest. If they are, the team will communicate
the research to the media and on our University website.

Following media coverage, concerns have been raised about the validity of
this research from academics in the fields of linguistics and medieval
studies. We take such concerns very seriously and have therefore removed
the story regarding this research from our website to seek further
validation and allow further discussions both internally and with the
journal concerned.

As I mentioned, I believe Cheshire's listing as a research associate has
also been removed from the UoB website.  To be clear, I don't get any
particular thrill from the idea of seeing this person, whatever his faults,
possibly under some sort of internal review.  The whole episode, with its
peer(less) review, journalistic blundering, etc., is simply like the
proverbial train wreck to me:  I can't help but stare.

> But then, standards are pretty low in the humanities, it seems,
> considering the recent boo-how about two evolutionary biologists
> (Russell Gray and Quentin Atkinson) "proving" the Anatolian hypothesis
> of Proto-Indo-European origins (which has long been *disproved* by
> linguists) in an article in the prestigious _Science_ journal, or the
> classic case of Alan Sokal fooling the journal _Social Text_ with an
> entirely meaningless article:
> > While looking into this, I find I am no longer able to locate the "who
> is"
> > web page for the author, Gerard Cheshire, on the University of Bristol
> > website.  I don't want to leap to an unwarranted conclusion, but I am
> sure
> > I saw it there.  From my memory, Cheshire was listed as a research
> > associate in evolutionary biology or something akin.
> So we are dealing with the usual case, as with Gray and Atkinson, where
> credentials in a "hard" science are abused to lend false credibility to
> crackpot work in the humanities.

Well, this time at least there is some hard back-pedaling, and not with
much delay, not only at UoB but also in the BBC and many other media
outlets that had picked up the story and run with it.  It's also been taken
down at (the link in John Q's OP).

> > At any rate, it is
> > clear that this paper is destined for inclusion in the long list of
> > rejected attempts, and not a very notable attempt at that, as
> Latin/Romance
> > decipherments have already been tried.
> Yes. As I wrote a few hours ago, while a hermetic text, like what the VM
> seems to be judging by the illustrations, would be expected to be
> written in encrypted Latin, the ciphers of the 15th century are so easy
> to break that if this was the case, the VM would long since have been
> deciphered, so this trace is cold and leads nowhere, and any effort at
> deciphering it that way is wasted.

Indeed.  Going by memory, according to Nick Pelling's cipher blog, we have
some idea how 15th century ciphers worked, and among others they sometimes
used the following methods to make covertext harder to crack:
- having more than one symbol for each vowel
- having special symbols for double letters
- having special symbols for common words
- adding meaningless "null" glyphs just to add confusion

All of these methods tend to grow the cipher alphabet.  But we know that
most of the Voynich text is written using 25 letters or so, depending how
you count.  So this does not appear to be a typical 15th century cipher,
partially because there are too few symbols.  On top of that, the Voynich
graphotactics simply do not seem to line up with known natural language
graphotactics; the Voynich graphotactics seem to be unnaturally
restrictive.  One can say, well, they used different symbols for a letter
in different positions, but that doesn't really work, because again, there
aren't enough letters for that kind of scheme.

> It is indeed more likely that it is
> written in an unknown language - either a lost natlang or a conlang. And
> it is perhaps not even encrypted, as the choice of the language alone
> makes it hard enough to crack!

It could be so!  I wouldn't be too surprised if that turned out to be the

When I was looking into this stuff some time (years) ago I tended to lean
toward the idea of a lost code book, where the Voynich tokens are not real
words but rather indices into code pages that would have listed either a
Latin word or words, or perhaps just a short sequence of Latin letters.  I
had the hunch that the "gallows" symbols, which often occur at the
beginning of a page or paragraph, were telling you what code page to use.
Somehow, the tokens were telling you where on the code page to find the
plain text, the plain-text bits perhaps being organized either in a simple
list or in a grid of rows and columns.  I believed that this effectively
indical/number-like encoding of the Voynich tokens would help explain some
of their statistical attributes (though not everything).  To be honest,
though, I would have to look up the statistical research paper that I was
reading at the time in order to completely remember why I was thinking this
way.  And admittedly my leaning towards this particular hypothesis was
largely a hunch on my part.  (I hope no one is bothered too much by me
indulging my hunch on this list.)