From gauss!rand.org!jim%mycroft Thu Dec 05 21:27:26 PST 1991
Received: by gauss; Fri Dec  6 00:29:50 EST 1991
Received: by inet.att.com; Fri Dec  6 00:29 EST 1991
Received: from mycroft.rand.org by rand.org; Thu, 5 Dec 91 21:27:31 -0800
Received: from localhost by mycroft.rand.org; Thu, 5 Dec 91 21:27:26 PST
Message-Id: <9112060527.AA09286@mycroft.rand.org>
To: voynich@rand.org
Subject: Testing...
Date: Thu, 05 Dec 91 21:27:26 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

This is a test of the Voynich Manuscript mailing list.

You're on it.

To get off it, send mail to voynich-request@rand.org .

To send mail to everybody on it (the two of us, so far), address it to
voynich@rand.org .

	Jim Gillogly

From gauss!gauss.att.com!reeds Fri Dec  6 11:07:45 EST 1991
Received: by gauss; Fri Dec  6 11:48:43 EST 1991
Received: by inet.att.com; Fri Dec  6 11:48 EST 1991
Received: from research.att.com by rand.org; Fri, 6 Dec 91 08:08:36 -0800
Message-Id: <9112061608.AA26697@rand.org>
From: reeds@gauss.att.com
Date: Fri, 6 Dec 91 11:07:45 EST
To: voynich@rand.org
Status: OR

Hi, Voynich-list-members!

1. Let me suggest we keep track of corrections to D'Imperio's 
transcription, as follows:

	Three files be kept at rand:  the original version
	of the transcription, the current best version, and
	a log of changes.  For ftp'ers sake, might also want
	to keep a current diff output.

	Lets say I find a typo.  I send mail to rand, saying 
	something like this:

-----
Here's an obvious correction, even though I'm reading fuzzy Kahn p865:
Blatent hando in line 15201, word 6: should be R not D.
Suggested revision:
15201B 4CPC89/ZC89/4OPOE/O8AE/SAE/8AR/4OBZC89/4OPC89/4OPC89/4OFC89/8AE-
-----

	Then the editor at rand makes the indicated change to the
	current version file, and saves my letter (possibly condensed)
	in the change log.

2. How long is the MS?  D'Imperio says 250,000 chars in several places.
But it seems to me that roughly half the pages are transcribed, and
when stripped of line numbers her transcription takes up 83,000 chars.
So are the untranscribed pages wordier?

3. A useful project for someone to do would be to make a list of
all of the pages, keeping track of the pagination with identifying 
notes like: 1r = Petersen p1, all the way through.  I, started to do
this from evidence in hand (a few plates, plus the pagination in
D'Imp's transcription), but cannot mesh it smoothly with D'Imp's 
Fig 12 on her p90. 

Of course any further info recorded per page would be all to the good.
(Text line count, brief description of marginalia, etc).  But a bald
list of actual pages is still lacking.

Regards, & Merry Xmas,

Jim Reeds
reeds@research.att.com
908 582 7066

From gauss!gauss.att.com!reeds Sat Dec  7 00:50:42 EST 1991
Received: by gauss; Sat Dec  7 01:08:27 EST 1991
Received: by inet.att.com; Sat Dec  7 01:08 EST 1991
Received: from research.att.com by rand.org; Fri, 6 Dec 91 21:51:18 -0800
Message-Id: <9112070551.AA11837@rand.org>
From: reeds@gauss.att.com
Date: Sat, 7 Dec 91 00:50:42 EST
To: voynich@rand.org
Status: OR

Jim,

I've sent off the letter to the library.  You are right: the expense
might be the limiting factor.

A colleague (Doug McIlroy) suggested that it would be more efficient
to redundantly retranscribe the MS and mechanically check for
discrepancies than it would be to transcribe once and proofread
redundantly.  At about one transcription goof per 100 characters,
(which is what we see with D'Imperio, alas) twofold redundancy cuts 
the goof rate to 1 per 10,000 chars, which would mean 20 in the whole 
book.  (This does not count actual substantive disagreements: just eye-
skip and handos.)  We toyed with the idea of ressurecting Bob Morris's 
old "typo" program, which trained itself on trigram stats, and spotted 
odd words, as a means of automatically detecting possible goofs.

I don't have any other changes to Currier to suggest, but a couple
of new characters might show up.  I thought I saw a new one in the
plate in Kahn, near p.845:  in the wedge between the two stems 
there is a pair of paragraphs of 6 and 3 lines.  The 2nd line of
the 2nd para reads something like 8AM 89'89 where 9' might be
a new character?

As for tricky pages:  I suppose in the end we just have to make a
diagram and whereever V text appears (be it a word, a line, or a para),
define a ``locus'', with locus identifier entered on the diagram,
and tag the transcribed text with page/locus/line-num.
Thus, on the page shown on Kahn p865, in addition to the usual
locus for lines (viz, the main body of text) we could define 8 more
loci, call them N1, N2, N3, N4, N5, W1, W2, and E1,
and have lines in the transcription like:

	152 N1 1	OFAN/AFOE	; ladies with hands in tubing
	152 N2 1	OPOE/ZC89	; under N1
	152 N3 1	OEFS8OE		; center top
	152 N4 1	OPOEOR
	152 N5 1	ORSC8AE		; under N4
	152 W1 1	2ORORAE		; above lady's head
	152 W2 1	OECOC8N		; on her vascular boat's hull
	152 E1 1	OFA

We might enter the approx. grid coords of the loci (accurate to a cm, say)
into the computer but I don't think we should attempt to represent anything
more complex about the geometric arrangement, or possible continuations
of text from one locus to another.  We should content ourselves with 
giving each ``line of text'' a unique name, and in accurately transcribing 
all lines of text.  (A line of text being any unbroken string of words.)

Nothing prevents us from putting descriptive comments into the
transcription.

I don't really know how many people we can round up.  A lot of people
would be game for a short time.  I think I could cajole 5 people where
I work into each transcribing 5 or so pages, but then they would stop.
I think you are right about netnews readers talking more than doing,
but you never know unless you try.  And we have all the time in the
world:  big V has been out there for a long time.

Jim

From gauss!rand.org!jim%mycroft Sat Dec 07 07:26:26 PST 1991
Received: by gauss; Sat Dec  7 10:28:40 EST 1991
Received: by inet.att.com; Sat Dec  7 10:28 EST 1991
Received: from mycroft.rand.org by rand.org; Sat, 7 Dec 91 07:26:36 -0800
Received: from localhost by mycroft.rand.org; Sat, 7 Dec 91 07:26:27 PST
Message-Id: <9112071526.AA00678@mycroft.rand.org>
To: reeds@gauss.att.com
Subject: Error checking and comments
Cc: voynich@rand.org, jim%mycroft@rand.org
In-Reply-To: Your message of Sat, 07 Dec 91 00:50:42 -0500.
             <9112070551.AA11837@rand.org> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Sat, 07 Dec 91 07:26:26 PST
Sender: jim%mycroft@rand.org
Status: OR

Jim -

Meta-comment:  Your messages don't include Subject: lines -- is that part
of your style, or is something eating them?  I find them critical to
finding an old message in reasonable time, so if they're getting munched
I'd like to find what's doing it. (If it's your style, I'll add my own
subject lines to them as they arrive here.)

Doug's idea of retranscribing the whole thing sounds on the whole better
than the proofing plan.  However, reading with an eyeball on both might
be helpful for spotting ambiguities (yeah, I guess I'd call that an a,
but maybe we should put an optional o on it as well).  For the first pass,
though, I think I agree.

> skip and handos.)  We toyed with the idea of ressurecting Bob Morris's 
> old "typo" program, which trained itself on trigram stats, and spotted 
> odd words, as a means of automatically detecting possible goofs.

Is that Bob Sr?  (joke alert coming) RAND has had enough Bob Jr code
running here to last us a few more years!  Seriously, looks like an
interesting idea.  Ideally we should train it separately on hand A and
hand B, which Currier says are closely linked to Language A and Language B.
For the ones with ambiguous hand or language (later pages) we ould try it
with both training sets.  Hmm.  Fun.

> I don't have any other changes to Currier to suggest, but a couple
> of new characters might show up.  I thought I saw a new one in the
> plate in Kahn, near p.845:  in the wedge between the two stems 
> there is a pair of paragraphs of 6 and 3 lines.  The 2nd line of
> the 2nd para reads something like 8AM 89'89 where 9' might be
> a new character?

That's page 64 (D'I) and 33v (V) for those of you following along in the
hymnals.  D'I calls that line between the stems /8AM/82/89- .  I'd have put
a dash before the 8AM also, since the line registration seems to change.

Looking at my slightly faded print, I'd say you're right.  There's a 2 on
the next line (off-screen from Kahn) that's totally different, and typical
of 2.

I have a few changes to suggest -- more for format than content:

1) The '/' doesn't do anything for me: I suggest we drop it and use spaces.
2) The comments you put on the "tricky" page are useful; we should have
   a comment format (the ';' to end of line convention seems fine) to
   use anywhere -- point out misregistration, long decorative ligatures,
   etc.
3) Let's drop the D'I page numbering and go to 1r, 33v and so on, using
   D'I's identification for the fold-out mentioned in the monograph rather
   than Newbold's.

Speaking of ligatures, I normally treat two separated S's joined with a
P as two Q's, figuring it for an entertainment by the copyist.

> As for tricky pages:  I suppose in the end we just have to make a

Looks like a good proposal.  "ladies" isn't PC, though -- "women"?

> 152 N1 1	OFAN/AFOE	; ladies with hands in tubing
> 152 N2 1	OPOE/ZC89	; under N1
> 152 N3 1	OEFS8OE		; center top
> 152 N4 1	OPOEOR
> 152 N5 1	ORSC8AE		; under N4

Should be:
  152 N4 1      OPOEOR          ; under N4
  152 N5 1      ORSC8AE         ; top right

> 152 W1 1	2ORORAE		; above lady's head

I'd be tempted to put

  152 W1 1      2O[2R]ORAE      ; above lady's head

> 152 W2 1	OECOC8N		; on her vascular boat's hull

I think the "CO" in this one is pretty common, and since it's joined,
maybe it should be a separate (new) character.  (?)

> We might enter the approx. grid coords of the loci (accurate to a cm, say)
> into the computer but I don't think we should attempt to represent anything
> more complex about the geometric arrangement, or possible continuations
> of text from one locus to another.  We should content ourselves with 
> giving each ``line of text'' a unique name, and in accurately transcribing 
> all lines of text.  (A line of text being any unbroken string of words.)

Agreed - if we don't keep referring back to the Ms we'll lose touch.

I'll post an invitation to join voynich@rand.org to sci.crypt.

Jim

From gauss!gauss.att.com!reeds Sat Dec  7 15:58:06 EST 1991
Received: by gauss; Sat Dec  7 16:30:58 EST 1991
Received: by inet.att.com; Sat Dec  7 16:30 EST 1991
Received: from research.att.com by rand.org; Sat, 7 Dec 91 12:58:23 -0800
Message-Id: <9112072058.AA24235@rand.org>
From: reeds@gauss.att.com
Date: Sat, 7 Dec 91 15:58:06 EST
To: voynich@rand.org
Status: OR

Jim -

Re Subject lines.  The mailer I use does not have Subject: lines.
(Its part of the minimalist philosophy at the part of the labs where 
I work.)  So treat it as my quirky personal style.  If it bugs you 
I could use a different mailer to send letters to you.

Side question:  on which syllable(s) do you put the stress when you
pronounce your surname?

Indeed Bob Sr.  You have been running a lot more of his code for a lot
longer than ever you ran of his kids code!

Substantivia:

1.  We should write a more detailed plan, suitable for handing to possible
contributors.  Needn't be super detailed, but should draw together all the
bits & snippets we've mentioned.

2.  Soon after we get a complete set of photocopies we should run off
a complete xerox of the whole thing.  High resolution and legibilty not
required.  I just want to have a overall view of the arrangement of
pages, pagination, layout of pages, stuff like that.  Maybe a reference
grid can be superimposed on this copy, folio numbers, ``loci'', and the 
like, should be written.  Print up 100 or so copies, call it ``An Atlas 
of the Voynich MS'', bind it with written plan described above, and give 
or sell copies to would-be transcribers.  Or get Aegean Park Press to do so.

3.  Typo program:  sure, I'll see what I can do.  We dug up the v6
source, fixed C anachronisms, got it to compile & run.  'typo<typo.c'
found words like 'argument' and 'perror', but 'typo<voynich' core
dumped.  Known hand A vs hand B training sets sounds good.

4.  Even with Doug's plan in effect, we will have to do a LOT of eye-
balling to resolve disputes.  AND will have to write a version comparer
which will cope with the fancy formats we come up with.  Its still
worth it, though.  Maybe if one transcription says POTAHTO and the other
says POTAYTO, the output of the comparison should say POTA<H,Y>TO,
and the human editor searches for <x,y> in the output.

5. About your format suggestions:  I agree with all your suggestions
(space sted /, comments at end of line, page numbering).  But I don't
understand D'I's identification of the parts of the fold-out, and probably
won't till I see the MS or photos.  (That's one thing I want the Atlas
for.)

6. When we discover new characters we should take care to give
them transcription values which do not conflict with Currier's names.
Digraphs or lower case, I suppose.  Maybe x1, x2, ... ?


7. I would consider ligatured CO as a separate char. only if we
could find pages showing both the ligatured and unligatured forms.
Why not 4O?  

8. sci.crypt: good.

Jim

From gauss!math.mit.edu!jbaez Sat Dec  7 16:40:36 EST 1991
Received: by gauss; Sat Dec  7 17:08:42 EST 1991
Received: by inet.att.com; Sat Dec  7 17:08 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Sat, 7 Dec 91 13:43:13 -0800
Received: from cayley (CAYLEY.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA10497; Sat, 7 Dec 91 16:39:31 EST
Received: by cayley; Sat, 7 Dec 91 16:40:36 EST
Date: Sat, 7 Dec 91 16:40:36 EST
From: jbaez@math.mit.edu
Message-Id: <9112072140.AA15347@cayley>
To: voynich@rand.org
Subject: Hi
Status: OR

Hi, I'm John Baez, who perhaps unwittingly catalyzed the formation of this
group.  I have sent to Yale in order to purchase a transcription of the 
Voynich ms... but perhaps this is unnecessary, if I can get ahold of it from
one of you?  

I would also like to know what the current plan is for transcribing the
thing.  

I am planning to visit the Beinecke collection at Yale and take a look
at the manuscript sometime in January.  (I am not yet 100% sure they'll
let me.)  If there are things people are seriously curious about that could
be answered by examination of the ms. I could try to find out.  I suspect
I won't have the patience to note the various colors of ink on all the 
illustrations... but I have a friend who will look at the ms. in the infrared.
I am trying to get some paleographical honchos aboard to try to persuade Yale
to permit a carbon-dating.

I don't know who most of you are so I'm a bit curious.

jb

From gauss!rand.org!jim%mycroft Sat Dec 07 14:07:40 PST 1991
Received: by gauss; Sat Dec  7 17:10:42 EST 1991
Received: by inet.att.com; Sat Dec  7 17:10 EST 1991
Received: from mycroft.rand.org by rand.org; Sat, 7 Dec 91 14:07:48 -0800
Received: from localhost by mycroft.rand.org; Sat, 7 Dec 91 14:07:40 PST
Message-Id: <9112072207.AA01470@mycroft.rand.org>
To: reeds@gauss.att.com
Subject: Voynich diffs
Cc: voynich@rand.org, jim%mycroft@rand.org
In-Reply-To: Your message of Sat, 07 Dec 91 15:58:06 -0500.
             <9112072058.AA24235@rand.org> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Sat, 07 Dec 91 14:07:40 PST
Sender: jim%mycroft@rand.org
Status: OR


> reeds@gauss.att.com writes:

> Side question:  on which syllable(s) do you put the stress when you
> pronounce your surname?

	gill OH glee  (all hard g's)

> 1.  We should write a more detailed plan, suitable for handing to possible
> contributors.  Needn't be super detailed, but should draw together all the
> bits & snippets we've mentioned.

Agreed.  I've got a column to write this weekend, but may have a chance
to draft something unless you can get to it first.

> 2.  Soon after we get a complete set of photocopies we should run off
> a complete xerox of the whole thing.  High resolution and legibilty not

Good plan -- Wonder if Aegean would be interested?  It's a lot of work to
muck around with multiple copies of that many pages.

> 4.  Even with Doug's plan in effect, we will have to do a LOT of eye-
> balling to resolve disputes.  AND will have to write a version comparer
> which will cope with the fancy formats we come up with.  Its still
> worth it, though.  Maybe if one transcription says POTAHTO and the other
> says POTAYTO, the output of the comparison should say POTA<H,Y>TO,
> and the human editor searches for <x,y> in the output.

I've got one of those -- a guy here at RAND (Walt Hobbs) put together a
Perl program to find small differences in output from one of the big fat
combat models we run.  It runs "diff", then marks the differing chars with
an underscore.  For example, this is the output so far of diffs from the
FTP version of the manuscript for the errors we've reported to each other.
The '_' doesn't stand out well with so many upper-case letters, so maybe
we should change it to a '*'.

                  _                 
65 00507A OPSOE/4O,AM/SOJ/ZOJ/8AJO-
65 00507A OPSOE/4O8AM/SOJ/ZOJ/8AJO-
                          _                                
939 07704B OR/AM/SCFO89/8A_-4OFOE/OFAM/OFAR/OF9/OFOE89/OE-
939 07704B OR/AM/SCFO89/8AR-4OFOE/OFAM/OFAR/OF9/OFOE89/OE-
                             _                         
1637 15226B 4OAR/ZC89/4OFAM/ZX9/4OPC89/RSR/OEPC89/ES9-
1637 15226B 4OAR/ZC89/4OFAM/ZQ9/4OPC89/RSR/OEPC89/ES9-
                                         _         
1648 15237B PSCC9/4OFCC9/4OFAE/ESC89/SCOEVAM/OFOE-
1648 15237B PSCC9/4OFCC9/4OFAE/ESC89/SCOEFAM/OFOE-


> 6. When we discover new characters we should take care to give
> them transcription values which do not conflict with Currier's names.
> Digraphs or lower case, I suppose.  Maybe x1, x2, ... ?

Be a shame to use digraphs, with everything else single chars.  I'd
prefer to head for lower case.

> 7. I would consider ligatured CO as a separate char. only if we
> could find pages showing both the ligatured and unligatured forms.

OK, I'll keep my eyes open -- couldn't find any other examples immediately.
Most CO are connected at the bottom, which is why this one stood out and I
thought it might deserve its own neo-Currier symbol.

> Why not 4O?  

Why not 40 for a separate char, you mean?  Why?  4 does appear as a separate
character fairly often.  (Would you believe about 77 times in this sample?)
Do you see different kinds of connections with it?

Jim Gillogly

From gauss!rand.org!jim%mycroft Sat Dec 07 14:33:18 PST 1991
Received: by gauss; Sat Dec  7 17:50:30 EST 1991
Received: by inet.att.com; Sat Dec  7 17:50 EST 1991
Received: from mycroft.rand.org by rand.org; Sat, 7 Dec 91 14:33:21 -0800
Received: from localhost by mycroft.rand.org; Sat, 7 Dec 91 14:33:19 PST
Message-Id: <9112072233.AA01503@mycroft.rand.org>
To: voynich@rand.org
Subject: Voynich availability and foliation
Date: Sat, 07 Dec 91 14:33:18 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

Jim Reeds has been keeping track of where different pieces of the text
have been published.  I have a copy of the British Museum microfilm
referred to in Kahn's references, but it's only up to f 55r, which doesn't
even make it through the "herbal" section (ff 1-11, 13-66 according to
Newbold p. 45).  I don't know that the BM doesn't have the rest, but
there's an END frame after this.  Jim, do you want to be the official
keeper of your list?  Do you want to keep the current version there for
ftp, or send it to me as you make changes to put up here at RAND?

I use the term "herbal" because of the illustrations.  Currier has
observed that the "language" used (on the basis of statistical observations)
varies from section to section, and doesn't exactly correspond to the
illustrations.  In particular, he says:

        The Newbold foliation indicates that the Biological Section
        extends through ff 85-86 and it would appear from the
        illustrations that the Pharmaceutical Section does not begin until
        f 87.  However, frequency counts before and after the break at f
        84/f 85 indicate a change from the Biological material to
        something else.  For example, the final "089", which does not
        occur in the Biol.  B text, shows up in ff 85-86 with quite a
        respectable frequency and matches the frequency of this final in
        the Pharmaceutical 'B' text of ff 94-95.

John asked who we are.  I'm a computer scientist currently at RAND, and
nobody pays me to do anything remotely connected with this.  I do
occasional papers for Cryptologia, am an editor for the American
Cryptogram Association, and have been working on the Voynich off and on
(mostly off) since the early 70's.  I consider myself (cringe) a scholar.

To find out who's currently on the list, I suppose a technically-oriented
network weenie's first idea would be to telnet to rand.org on port 25 and
do a "vrfy voynich" and then a "quit".  There are four of us as I write,
but I expect more to show up as the invitation propagates through the Net.

	Jim Gillogly

From gauss!gauss.att.com!reeds Sat Dec  7 18:31:22 EST 1991
Received: by gauss; Sat Dec  7 19:08:42 EST 1991
Received: by inet.att.com; Sat Dec  7 19:08 EST 1991
Received: from research.att.com by rand.org; Sat, 7 Dec 91 15:31:34 -0800
Message-Id: <9112072331.AA26825@rand.org>
From: reeds@gauss.att.com
Date: Sat, 7 Dec 91 18:31:22 EST
To: voynich@rand.org
Status: OR

Hi, VMS fans!

This is inresponse to jbaez@math.mit.rand 's recent letter
to voynich@rand.org:

> Hi, I'm John Baez, who perhaps unwittingly catalyzed the formation of this
> group.

Indeed you did!  Many thanks!  

John wrote

> I don't know who most of you are so I'm a bit curious.

I am Jim Reeds, work in the Math Research Center at AT&T Bell Labs,
became first interested in the Voynich MS about 25 years ago, found
my interest rekindled by news that Jim Gillogly has a machine readable
transcription of about half of the MS.

Jim Gillogly and I have been talking about making a complete transcription,
fulfilling the wish of D'Imperio, expressed in section 11.4, para 2, page 78
of her book,  cheking for accuracy, etc.  This depends on having a full
photgraphic copy of the MS.

We have spent a lot of time discussing ways and means.  Maybe you should
read the backlog of letters we have sent?  (Jim, can you make them ftp-able?)
In the mean time Jim or I will try to write down our ideas in more systematic
form.  If we go ahead, a lot of hard tedious work is involved, and the
more help we can get the better.  

Jim Reeds
reeds@research.att.com
908 582 7066



From gauss!rand.org!jim%mycroft Sat Dec 07 17:29:12 PST 1991
Received: by gauss; Sat Dec  7 21:11:09 EST 1991
Received: by inet.att.com; Sat Dec  7 21:11 EST 1991
Received: from mycroft.rand.org by rand.org; Sat, 7 Dec 91 17:29:15 -0800
Received: from localhost by mycroft.rand.org; Sat, 7 Dec 91 17:29:12 PST
Message-Id: <9112080129.AA02202@mycroft.rand.org>
To: voynich@rand.org
Subject: Voynich: Proposed welcome message
Date: Sat, 07 Dec 91 17:29:12 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

Here's a proposed welcome message for new list members.  Please suggest
additions and corrections...

I've put up a digest of the initial (and a few pre-initial) messages
that got this stuff started.  Should it be in some standard digest format?

	Jim Gillogly

------

Welcome to the Voynich Manuscript mailing list.

Send mail to voynich-request@rand.org to get off of it again or if you're
having distribution problems or problems with the FTP archives.

Send substantive mail to voynich@rand.org and it'll get to the whole group.

A transcription by Mary D'Imperio of a good chunk of the Manuscript is
available for anonymous ftp at

	Host:   rand.org  (192.5.14.33)
	Dir:    pub/jim
	File:   voynich.tar.Z

As time permits, a digest of the mailing list will be in this directory
under some obvious name.  It'll be useful reading to catch you up on
the plans.

As we gather other materials, they'll go into the same directory.  This
will include a bibliography (starting with John Baez' posting), and
Voynich fonts.  Jim Reeds (reeds@gauss.att.com) is maintaining a list of
publicly accessible reproductions of pages of the Voynich, including
rationalizing the foliation used by different sources.

The plan is to use this mailing list to coordinate a new investigation of
the Manuscript.  For this plan to achieve anything useful, we need to
agree on some ground rules.  If you have a problem with these ideas or
have spotted a bogosity, SPEAK UP SOON!  The time to apply leverage for
course corrections is at the beginning of the project!

The current proposal:

(1) We will try to get one or more high-quality copies of the Voynich
    Manuscript from the Beinecke Rare Book Library, probably on microfilm.

(2) We will make lower-quality "road map" copies of it available by some
    means yet to be determined -- perhaps encouraging Aegean Park Press
    to print it.

(3) If we can find earlier transcriptions, we will get them and convert
    to a machine-readable format.

(4) The next order of business is to complete the transcription --
    the D'Imperio transcription is about half of the 116 folios.
    (That's 131 out of a total of 262 pages, by Newbold's count.)
    We propose to use an extension/modification of Currier's notation,
    including comments, replacing slashes with spaces, and including
    captions on pictures and ambiguous readings.

(5) Jim Gillogly (jim@rand.org) has volunteered to coordinate the
    transcription effort, keeping a current "best guess" in the FTP
    directory with diffs from the D'Imperio transcription.  This will
    involve farming out high-quality text copies of selected pages to
    volunteers.  Ideally we will have three people transcribe each page
    and compare the results... this depends on the labor available, of
    course.

From gauss!rand.org!jim%mycroft Sat Dec 07 19:57:02 PST 1991
Received: by gauss; Sat Dec  7 23:08:32 EST 1991
Received: by inet.att.com; Sat Dec  7 23:08 EST 1991
Received: from mycroft.rand.org by rand.org; Sat, 7 Dec 91 19:57:05 -0800
Received: from localhost by mycroft.rand.org; Sat, 7 Dec 91 19:57:02 PST
Message-Id: <9112080357.AA02507@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: ground rules proposal: comments 
In-Reply-To: Your message of Sun, 08 Dec 91 13:51:35 -0500.
             <9112080251.AA07933@medici.trl.OZ.AU> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Sat, 07 Dec 91 19:57:02 PST
Sender: jim%mycroft@rand.org
Status: OR


> j.guy@trl.OZ.AU (Jacques Guy) writes various good stuff:
> 
> Yes.  Did my short msg about scanning it all into GIF get through the
> net?  More fore my reasons to propose that later.

Yes, it got here -- more on my reasons to be nervous later.

> (3) If we can find earlier transcriptions, we will get them and convert
> to a machine-readable format.
> 
> Mmm, I wonder.  How?  Typing them in or by OCR?  Earlier
> transcriptions are not necessarily better, or even accurate.  For
> instance, Levitov inserts letters where it helps his thesis.

Yuck!  We were thinking of trying to get the some of the materials
mentioned by D'Imperio; it'd be nice to get them machine-readable, but, as
you say, OCR would be preferable.  Working with text produced by Friedman
and Tiltman, for example, would at least have a lot of relevant experience
attached... only exceeded by having a linguist on the team, I hasten to
add.

> Goyoojin again!  Currier's notation strikes me as pretty good.  Let's
> introduce innovations only when needed.  A protocol for comments would
> be *very* nice indeed.

Of *course* we'd only be making the necessary improvements!  Comments,
yes.  Everybody agree with "Semicolon initiates comment to end of line"?

Another potential change is calling J and 7 the same character.  Can
anybody tell the difference?

There are characters in there that don't have Currier equivalents, such
as 9' that Jim Reeds pointed out.  They'd need some convention.

>                         Replacing slashes with spaces?  I'd say no.
> Spaces are useful for pretty-printing bilingual texts with interlinear
> translations.  I would keep slashes (or any other printing symbol) for
> word-delimiters.  It's the linguist in me speaking here, with just a
> seat-of-the-pants feeling, difficult to explain in a reasoned
> argument, just a gut-feeling.

I guess.  The space-to-slash conversion I would expect to be straightforward:
a one-line Perl script should be able to go either way.  For my own use I
prefer to look at it without the spaces, but I wouldn't mind running my
Perl program on it first.

> manuscript is this problem: given a text in a unknown script, find the
> alphabet of the script.  In other words, it's the long standing

Right -- that's the story behind the differences among the alphabets of
each of the different study groups.  We don't know how to segment it until
we've cracked it.  D'Imperio's alphabet (as opposed to Currier's) was
designed to leave those questions unanswered: building things out of the
O-ligature and the ligature-O, for example, and having the \ as a separate
character, as opposed to Currier's EGH1 set, RTU0 set, and DNM3 set, which
in this script seem to have lost the identity of their components.  Another
reason we need to keep in touch with the Manuscript rather than working
exclusively with the transcription.

> here to a digitized image.  Example: find the word breaks in:
> atheungeunte've'ghriyamayanlethemossoneshovrunethemlyanmlyanmeugheuth

Point taken.  Dass schlagt von mir die Scheisse heraus, as we might say.
I've been proofing some of D'I's early pages, and finding that I often
disagree about whether there was a word break.

> Now you can see where I got my idea of scanning the Voynich into
> GIF images.

Yes, interlinear work would be wonderful -- I used it constantly while
breaking an 18th century shorthand system, when (again) I didn't know
where one character started and another began.  I did a bunch of xeroxing
and paste-up, and it took forever dealing with just one page.  A
heavy-duty image/text system would be great.

How would we deal with the volume of images?  A reasonably leerable GIF of
the form that seems to be bringing the net to its knees runs about 200K.
We'd probably want more pixels for the kind of work we're talking about --
say 400K?  For 262 pages (don't know whether that includes the fold-out),
that comes to (I can do this, just a second) about 100 MB.  I can probably
stretch the FTP area on Randvax to about 1 MB, but we certainly don't have
a spare 100 MB there.  Would we send them around on mag tapes?  Probably
shipping 100 PC diskettes wouldn't make it.  Is there a dramatically more
efficient format than GIF yet?

Jim Gillogly

From gauss!rand.org!jim%mycroft Sat Dec 07 23:15:11 PST 1991
Received: by gauss; Sun Dec  8 02:28:35 EST 1991
Received: by inet.att.com; Sun Dec  8 02:28 EST 1991
Received: from mycroft.rand.org by rand.org; Sat, 7 Dec 91 23:15:13 -0800
Received: from localhost by mycroft.rand.org; Sat, 7 Dec 91 23:15:11 PST
Message-Id: <9112080715.AA02952@mycroft.rand.org>
To: voynich@rand.org
Subject: Transcribed f 65v
Date: Sat, 07 Dec 91 23:15:11 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

Here's a possible format for new additions.  We've discussed going back to
folio designations (recto and verso) rather than sticking with the page
numbers, which are less immediately transparent.  Here's a proposed reading
of f 65v, which is Plate II of Newbold (opposite page 44).  Are there two
volunteers to transcribe it without paying close attention to mine?

Comments welcome on the format... note the proposed [49] on line 6.  I
believe the first * and the [49] could be resolved with a clearer copy.

	Jim Gillogly
-------------------------
; Folio 65v (Herbal) appears as Plate II in Newbold.
; It features a plant with many roots, branches and flowers; the main
; stem separates the lines of text, which are well-registered across
; the stem.
; Transcribed 7 Dec 91 by Jim Gillogly.
;
065V01 W9/VSCY9/89/-/8SCBAN/ZCP9/4OB9/VOE/SB89-
065V02 8AM/ZCCF/E/O89/-/9PCO/4OB*/AT/SCOPCC9/8AEAJ-  ; * may be R
065V03 PAE/SCOP/ZF9/9/-/*AR/SCO89-                   ; * may be S or R or 2??
065V04 POCC89/OPOR/ZC89/-/9PCC9/ZCO89/4OFZ8/9VSC9-
065V05 8ZC89/OFCO89/4OF8/-/ZX9/SOF9/SOFCO89/OFC9/89-
065V06 [49]O/9PSC9/ZXCW9/-/9FSC89/SC89/ZX89#         ; Probably 4 (?)
-------------------------

From gauss!trl.OZ.AU!j.guy Sun Dec  8 11:04:42 EST 1991
Received: by gauss; Sat Dec  7 19:09:24 EST 1991
Received: by inet.att.com; Sat Dec  7 19:09 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sat, 7 Dec 91 16:04:56 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA07832; Sun, 8 Dec 91 11:04:42 EST
Date: Sun, 8 Dec 91 11:04:42 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112080004.AA07832@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: How about scanning the Voynich into GIF?
Status: OR


Well, the subject title says it all. Your chit-chat out there makes me green with envy. There isn't a single reproduction of the Voynich in the whole width and breadth of Australia. And Kahn's "Codebreakers" is out on loan, and the borrower is a away until after Xmas and....the gun laws are too strict so I can't even shoot myself in despair!

From gauss!trl.OZ.AU!j.guy Sun Dec  8 12:31:39 EST 1991
Received: by gauss; Sat Dec  7 21:09:42 EST 1991
Received: by inet.att.com; Sat Dec  7 21:09 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sat, 7 Dec 91 17:31:57 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA07893; Sun, 8 Dec 91 12:31:39 EST
Date: Sun, 8 Dec 91 12:31:39 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112080131.AA07893@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: self-introduction time
Status: OR


Prompted by John Baez question (QRU).

Jacques Guy.  Artificial Intelligence Systems, Research Laboratories of Telecom Australia, in Clayton, a suburb of Melbourne, 20 km SE of the City. Uni: Chinese, Japanese, and Tahitian at the Ecole des Langues Orientales in Paris, then, after an employment, unemployment, migrating, employment interlude, did a Ph.D. on a dialect of Espiritu Santo (New Hebrides, now Vanuatu). Spent 15 more years in the Linguistics Dept (Research School) of the Australian National University, eventually mostly working on methods of numerical taxonomy for reconstructing the prehistory of language families. With Telecom Australia since December 1985. Sorry, can't tell what I get paid for working on.  

Always wanted to be a sort of discoverer when I was a kid (or an inventor: for a while, I wanted to be Gyro Gearloose when I grew up!). Hence my interest in trying to decipher the Voynich. I have no illusions, though, on that count,  so I put more effort into the Easter Island tablets: those, at least, I am sure are not a hoax!

From gauss!trl.OZ.AU!j.guy Sun Dec  8 13:51:35 EST 1991
Received: by gauss; Sat Dec  7 22:37:46 EST 1991
Received: by inet.att.com; Sat Dec  7 22:37 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sat, 7 Dec 91 18:51:47 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA07933; Sun, 8 Dec 91 13:51:35 EST
Date: Sun, 8 Dec 91 13:51:35 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112080251.AA07933@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: ground rules proposal: comments
Status: OR


The current proposal:

(1) We will try to get one or more high-quality copies of the Voynich
    Manuscript from the Beinecke Rare Book Library, probably on microfilm.

    Yes, the microfilm I had was mediocre, with many frames right out of focus. 
    Colour would help disambiguate those case where a line from an illustration
    interferes with the text.

(2) We will make lower-quality "road map" copies of it available by some
    means yet to be determined -- perhaps encouraging Aegean Park Press
    to print it.

    Yes. Did my short msg about scanning it all into GIF get through the net?
    More fore my reasons to propose that later.
    

(3) If we can find earlier transcriptions, we will get them and convert
    to a machine-readable format.
    
    Mmm, I wonder. How? Typing them in or by OCR? Earlier transcriptions are
    not necessarily better, or even accurate. For instance, Levitov inserts letters
    where it helps his thesis. Careful there (goyoojin, goyoojin as some general 
    says in "Ran" after delivering a tongue-in-cheek learned lecture on werefoxes)
    

(4) The next order of business is to complete the transcription --
    the D'Imperio transcription is about half of the 116 folios.
    (That's 131 out of a total of 262 pages, by Newbold's count.)
    We propose to use an extension/modification of Currier's notation,
    including comments, replacing slashes with spaces, and including
    captions on pictures and ambiguous readings.

    Goyoojin again! Currier's notation strikes me as pretty good. Let's introduce
    innovations only when needed. A protocol for comments would be *very* nice indeed.
    Replacing slashes with spaces? I'd say no. Spaces are useful for pretty-printing
    bilingual texts with interlinear translations. I would keep slashes (or any other
    printing symbol) for word-delimiters. It's the linguist in me speaking here,
    with just a seat-of-the-pants feeling, difficult to explain in a reasoned 
    argument, just a gut-feeling. 


(5) Jim Gillogly (jim@rand.org) has volunteered to coordinate the
    transcription effort, keeping a current "best guess" in the FTP
    directory with diffs from the D'Imperio transcription.  This will
    involve farming out high-quality text copies of selected pages to
    volunteers.  Ideally we will have three people transcribe each page
    and compare the results... this depends on the labor available, of
    course.

    One of the reasons why I am interested in working on the Voynich manuscript is
    this problem: given a text in a unknown script, find the alphabet of the script.
    In other words, it's the long standing question of segmenting a continuous text
    into its components, applied here to a digitized image. Example: find the word
    breaks in:  atheungeunte've'ghriyamayanlethemossoneshovrunethemlyanmlyanmeugheuth
    No, you have no dictionary to rely on, no grammar, no-one to help you. Now, 
    given the above line of gibberish (it's not, it's a real language, and it's the
    beginning of a folk tale but never mind), imagine that you have forgotten the Roman
    alphabet and that you are staring at what it really is: an array of pixels. The
    problem now is: group those pixel formations into letters, find the alphabet.
    
    Now you can see where I got my idea of scanning the Voynich into GIF images.

    There would be an immediate advantage: with a proper image/word processor,
    the transcription chore would be made *IMMENSELY* easier: just insert a 
    blank line below every line of reproduction from the Voynich, and type in
    your transliteration. (I believe Wordstar 6 would allow you to do that, 
    judging from a demo disk I got out of a British magazine).


From gauss!trl.OZ.AU!j.guy Sun Dec  8 14:18:53 EST 1991
Received: by gauss; Sat Dec  7 22:38:27 EST 1991
Received: by inet.att.com; Sat Dec  7 22:38 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sat, 7 Dec 91 19:19:04 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA07947; Sun, 8 Dec 91 14:18:53 EST
Date: Sun, 8 Dec 91 14:18:53 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112080318.AA07947@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Vowel recognition and table frequencies: my source code (Turbo Pascal)
Status: OR


This is the source code of the program I wrote for computing the vowels and letter frequency tables of my article in Cryptologia. I used Turbo Pascal 5.5, but without resorting to objects. It ought to run unmodified in versions 4.0, 5.0, and 6.0 too, and with very minor modifications in version 3.0. It was a slap-bang whack-it-together-quick affair, so don't expect much of a user interface, abundant comments, and a programming style Edsgar Dijkstra would give me a summa cum laude for.

If you read through the code you'll see that it handles any line that starts with a semicolon as a comment (there are better ways of course of allowing for comments in the text, but as I said, that was a quick-and-dirty job).
Here we go:
----------------CUT HERE------------------------------
Uses Crt,Dos;
TYPE SetOfChar=SET OF CHAR; s80= STRING[80]; s20=STRING[20];
CONST Alphabet: SetOfChar=['A'..'Z','a'..'z'];
      NoChar=#0;
      MaxLetters=100;
VAR Uncapitalize: BOOLEAN;
    CodeOf: ARRAY[#0..#255] OF INTEGER;
    LetterNo: ARRAY[1..MaxLetters] OF CHAR;
    Fq,FqInitial,FqMedial,FqFinal,FqAlone,FqLineInitial,FqLineFinal,
    Vowel: ARRAY[1..MaxLetters] OF INTEGER;
    Cell: ARRAY[1..MaxLetters,1..MaxLetters] OF INTEGER;
    LastLetter: INTEGER; c: CHAR;
    SumAll,SumInitial,SumFinal,SumMedial,SumAlone,SumLi,SumLineFinal,
    Vowels: INTEGER;

FUNCTION NameOf(VAR f: TEXT): STRING;
VAR i: INTEGER;
BEGIN i:=1;
  WITH TextRec(f) DO
  WHILE Name[i]>' ' DO Inc(i);
  NameOf:=Copy(TextRec(f).Name,1,i)
END;

PROCEDURE Zero;
BEGIN
    FillChar(CodeOf,SizeOf(CodeOf),0);
    FillChar(LetterNo,SizeOf(LetterNo),0);
    FillChar(Fq,SizeOf(Fq),0);
    FillChar(FqInitial,SizeOf(FqInitial),0);
    FillChar(FqMedial,SizeOf(FqMedial),0);
    FillChar(FqFinal,SizeOf(FqFinal),0);
    FillChar(FqAlone,SizeOf(FqAlone),0);
    FillChar(FqLineInitial,SizeOf(FqLineInitial),0);
    FillChar(FqLineFinal,SizeOf(FqLineFinal),0);
    FillChar(Vowel,SizeOf(Vowel),0);
    FillChar(Cell,SizeOf(Cell),0);
    LastLetter:=0;
    LastLetter:=0;
    SumAll:=0;
    SumInitial:=0;
    SumFinal:=0;
    SumMedial:=0;
    SumAlone:=0;
    SumLi:=0;
    SumLineFinal:=0;
    Vowels:=0;
END;

FUNCTION Yes(Prompt: STRING): BOOLEAN; VAR c: CHAR;
BEGIN Write(Prompt);
   REPEAT c:=ReadKey; c:=Upcase(c) UNTIL c IN ['Y','N'];
   Yes:=c='Y';
END;

FUNCTION UnCaps(s: s80): s80; VAR i: INTEGER; c: CHAR;
BEGIN FOR i:=1 TO Length(s) DO
  BEGIN c:=s[i]; IF (c>='A') AND (c<='Z') THEN s[i]:=CHR(ORD(c)+32)
  END;
  UnCaps:=s
END;

PROCEDURE wait; VAR c: CHAR;
BEGIN GotoXY(1,24);
   write('Press any key to continue '); c:=ReadKey;
END;

FUNCTION GetFile(x,y: BYTE; Prompt: STRING; VAR f: TEXT): BOOLEAN;
(* the first character of the prompt must be         *)
(* 'i' or 'I' for an input file                      *)
(* 'o' or 'O' for an output file                     *)
VAR fn: s80; ok,found: BOOLEAN;
BEGIN GetFile:=FALSE; fn:=''; Reset(input);
  REPEAT Gotoxy(x,y); ClrEol; write(Prompt); read(fn);
    IF fn='' THEN Exit;
    Assign(f,fn); {$I-} Reset(f) {$I+}; found:=IOResult=0;
    CASE Prompt[1] OF
      'i','I': {Input}
             IF found THEN ok:=TRUE
             ELSE BEGIN ok:=FALSE; Gotoxy(x,y); ClrEol;
                write(fn+' not found. '+Prompt);
             END;
      'o','O': {Output}
             IF found
             THEN BEGIN Gotoxy(x,y); ClrEol;
                    ok:=Yes('Overwrite '+fn+'? (Y/N) ');
                  END
             ELSE ok:=TRUE;
    END
  UNTIL ok;
  GetFile:=ok;
END;


PROCEDURE CountThem(VAR f: TEXT; VAR n: INTEGER);
VAR s: string;
    i,k,PreviousLetter,LastChar,SecondLastChar,EndOfLine: INTEGER;
    c: CHAR;
    GotIt: BOOLEAN;
BEGIN PreviousLetter:=0;
  reset(f);
  WHILE NOT Eof(f) DO
  BEGIN Readln(f,s);
    IF s<>'' THEN IF s[1]<>';' {which shows a comment} THEN
    BEGIN
    IF Uncapitalize THEN s:=UnCaps(s); s:=s+' ';
    LastChar:=0; SecondLastChar:=0;
    FOR i:=1 TO Length(s) DO
    BEGIN c:=s[i];
      IF c IN Alphabet THEN
      BEGIN
        IF CodeOf[c]=0 THEN
        BEGIN n:=n+1; CodeOf[c]:=n; LetterNo[n]:=c;
        END;
        k:=CodeOf[c]; Fq[k]:=Fq[k]+1;
        IF PreviousLetter<>0 THEN
        BEGIN Cell[k,PreviousLetter]:=Cell[k,PreviousLetter]+1;
          Cell[PreviousLetter,k]:=Cell[PreviousLetter,k]+1;
        END;
        PreviousLetter:=k
      END;
      IF c IN Alphabet THEN
      BEGIN
        IF LastChar=0 THEN
        BEGIN FqInitial[k]:=FqInitial[k]+1
        END;
        SecondLastChar:=LastChar;
        LastChar:=k;
      END
      ELSE
      BEGIN
         IF LastChar<>0 THEN
         BEGIN
           IF SecondLastChar<>0 THEN FqFinal[LastChar]:=FqFinal[LastChar]+1
           ELSE
           BEGIN FqAlone[LastChar]:=FqAlone[LastChar]+1;
                 FqInitial[LastChar]:=FqInitial[LastChar]-1
           END
         END;
         SecondLastChar:=LastChar;
         LastChar:=0
      END
    END;
    EndOfLine:=Length(s);
    i:=0; GotIt:=false;
    REPEAT i:=i+1;
      c:=s[i];
      IF c IN Alphabet THEN
      BEGIN k:=CodeOf[c]; FqLineInitial[k]:=FqLineInitial[k]+1;
        GotIt:=true
      END
    UNTIL GotIt OR (i=EndOfLine);
    i:=EndOfLine; GotIt:=false;
    REPEAT i:=i-1;
      c:=s[i];
      IF c IN Alphabet THEN
      BEGIN k:=CodeOf[c]; FqLineFinal[k]:=FqLineFinal[k]+1;
        GotIt:=true
      END
    UNTIL GotIt OR (i=1);
    END;
  END;
  Close(f)
END;
PROCEDURE Sort; (* this is a Shell sort *)
VAR i,j,k,m,tmp,n: INTEGER; swap: BOOLEAN; s: s20; c: CHAR;
BEGIN n:=LastLetter;
  m:=1; WHILE m<=n DO m:=m+m; m:=(m-1) DIV 2;
  WHILE m>0 DO
  BEGIN FOR j:=1 TO n-m DO
    BEGIN i:=j; k:=i+m;
      REPEAT swap:=Fq[k]>Fq[i];
        IF swap
        THEN BEGIN
           tmp:=Fq[k]; Fq[k]:=Fq[i]; Fq[i]:=tmp;
           tmp:=FqInitial[k]; FqInitial[k]:=FqInitial[i]; FqInitial[i]:=tmp;
           tmp:=FqFinal[k]; FqFinal[k]:=FqFinal[i]; FqFinal[i]:=tmp;
           tmp:=FqAlone[k]; FqAlone[k]:=FqAlone[i]; FqAlone[i]:=tmp;
           tmp:=FqLineFinal[k]; FqLineFinal[k]:=FqLineFinal[i];  FqLineFinal[i]:=tmp;
           tmp:=FqLineInitial[k]; FqLineInitial[k]:=FqLineInitial[i]; FqLineInitial[i]:=tmp;
           tmp:=Vowel[k]; Vowel[k]:=Vowel[i]; Vowel[i]:=tmp;
           c:=LetterNo[k]; LetterNo[k]:=LetterNo[i]; LetterNo[i]:=c;
           i:=i-m; k:=i+m;
        END
      UNTIL (i<=0) OR NOT swap
    END;
    m:=m DIV 2
  END;
END;

PROCEDURE Calculate;
VAR i,n,
      FqAll,FqFin,FqInit,FqMed,FqLone,FqLI,FqLF: INTEGER;
BEGIN n:=LastLetter;
    for i:=1 to n do
    begin FqAll:=Fq[i]; FqInit:=FqInitial[i]; FqFin:=FqFinal[i];
        FqLone:=FqAlone[i];
        FqMed:=FqAll-FqInit-FqFin-FqLone;
        FqMedial[i]:=FqMed;
        SumAll:=SumAll+FqAll;
        SumInitial:=SumInitial+FqInit;
        SumFinal:=SumFinal+FqFin;
        SumMedial:=SumMedial+FqMed;
        SumAlone:=SumAlone+FqLone;
        SumLi:=SumLi+FqLineInitial[i];
        SumLineFinal:=SumLineFinal+FqLineFinal[i];
        if Vowel[i]>0 then Vowels:=Vowels+1;
    end
END;


(** Now here is the "meat" of Sukhotin's vowel algorithm, the bit that finds the 
    vowels from an existing matrix of letter adjacencies **)
{*****
(1) Fill main diagonal with zeroes.
(2) Let Sum(i) be the sum of the entries of the ith row.
    Calculate Sum(i) for each row.
(3) Let Cat(i) be the category (vowel or consonant) of letter i.
    Set all letters to consonant, (i.e. for i:=1 to n do Cat(i):=consonant).
(4) Select the letter m for which Sum(m) is maximum and Cat(i) is
    consonant.
    If Sum(m)> 0 then
       a) set Cat(m) to vowel,
       b) let Sum(i) = Sum(i)- f(i,m)*2 for all i's for which Cat(i)
          is consonant.
    Repeat until Sum(m)=0.
*****}
PROCEDURE FindVowels;
  VAR x,y,n, max,sum,ymax,ThisVowel: INTEGER; Stop: BOOLEAN;
  SumOf: ARRAY[1..MaxLetters] OF INTEGER;
BEGIN ThisVowel:=0;
  n:=LastLetter;
  FOR x:=1 TO n DO Cell[x,x]:=0;
  FOR y:=1 TO n DO
  BEGIN Sum:=0;
    FOR x:=1 TO n DO Sum:=Sum+Cell[y,x];
    SumOf[y]:=Sum
  END;
  REPEAT  max:=0;
    FOR y:=1 TO n DO IF (Vowel[y]=0) AND (SumOf[y]>max)
    THEN BEGIN ymax:=y; max:=SumOf[y] END;
    Stop:=max<=1;
    IF NOT Stop
    THEN BEGIN ThisVowel:=ThisVowel+1;
      Vowel[ymax]:=ThisVowel;
      FOR y:=1 TO n DO IF (Vowel[y]=0)
      THEN SumOf[y]:=SumOf[y]-Cell[ymax,y]-Cell[y,ymax];
    END
  UNTIL Stop
END;

PROCEDURE CapitalizeOrNot; VAR c: CHAR;
BEGIN
  Uncapitalize:=Yes('Change all uppercase to lowercase? (Y/N) '); writeln;
END;

PROCEDURE GetAlphabet;
VAR ChangeIt: BOOLEAN; Buffer: STRING[80]; c: CHAR; i: INTEGER;
BEGIN
  REPEAT writeln('Alphabet is now:');
    FOR c:=#33 TO #126 DO IF c IN Alphabet THEN write(c);
    writeln;
    ChangeIt:=Yes('Any changes? (Y/N) '); writeln;
    IF ChangeIt THEN
    BEGIN write('Remove: '); readln(buffer);
      FOR i:=1 TO Length(buffer) DO
      BEGIN c:=buffer[i]; Alphabet:=Alphabet-[c]
      END;
      IF buffer<>'' THEN
      BEGIN writeln('Alphabet is now:');
        FOR c:=#33 TO #126 DO IF c IN Alphabet THEN write(c); writeln;
      END;
      write('Add to it: '); readln(buffer);
      FOR i:=1 TO Length(buffer) DO
      BEGIN c:=buffer[i];
        IF (c>' ') AND (c<=#126) THEN Alphabet:=Alphabet+[c]
      END
    END
  UNTIL NOT ChangeIt
END;

PROCEDURE ClearSums;
BEGIN
  LastLetter:=0;
  FillChar(CodeOf,SizeOf(CodeOf),0);
  FillChar(Fq,SizeOf(Fq),0);
  FillChar(FqInitial,SizeOf(FqInitial),0);
  FillChar(FqFinal,SizeOf(FqFinal),0);
  FillChar(FqAlone,SizeOf(FqAlone),0);
  FillChar(FqLineInitial,SizeOf(FqLineInitial),0);
  FillChar(FqLineFinal,SizeOf(FqLineFinal),0);
  FillChar(Cell,SizeOf(Cell),0);
END;

PROCEDURE Title(VAR f: TEXT); (* prints the headings of the tables *)
BEGIN
    writeln(f,SumAll,' letters. ',Vowels,' vowels, ',LastLetter-Vowels,' consonants');
    writeln(f);
    writeln(f,
    '        Absolute frequency                 Relative Frequency (per 1000)');
    writeln(f,
'V#  Total Init  Med  Fin Isol  L/I  L/F    Total Init  Med  Fin Isol  L/I  L/F');
END;


PROCEDURE ShowLetterNo(VAR f: TEXT; i: INTEGER); (* prints the statistics of the i-th letter into file f *)

  FUNCTION RelFrq(n,Sum: INTEGER): INTEGER;
  BEGIN IF Sum=0 THEN RelFrq:=0 ELSE
       RelFrq:=Round(n/Sum*1000)
  END;

BEGIN
    IF Vowel[i]=0 THEN write(f,'  ') ELSE write(f,Vowel[i]:2);
    write(f,LetterNo[i]:2);
    writeln(f,Fq[i]:5,
          FqInitial[i]:5,
          FqMedial[i]:5,
          FqFinal[i]:5,
          FqAlone[i]:5,
          FqLineInitial[i]:5,
          FqLineFinal[i]:5,
          LetterNo[i]:4,
          RelFrq(Fq[i],SumAll):5,
          RelFrq(FqInitial[i],SumInitial):5,
          RelFrq(FqMedial[i],SumMedial):5,
          RelFrq(FqFinal[i],SumFinal):5,
          RelFrq(FqAlone[i],SumAlone):5,
          RelFrq(FqLineInitial[i],SumLi):5,
          RelFrq(FqLineFinal[i],SumLineFinal):5
          );
END;

PROCEDURE ShowResults(n: INTEGER);
CONST BottomRow=22;
VAR i,y: INTEGER;
BEGIN
  REPEAT i:=0;
    REPEAT ClrScr; Title(output); y:=4;
      REPEAT Inc(i); Inc(y); ShowLetterNo(output,i);
      UNTIL (y=BottomRow) OR (i=n);
      IF i<n THEN
      BEGIN GotoXY(1,24); write('--- There is more  (press any key) ---');
        REPEAT UNTIL KeyPressed;
      END;
    UNTIL i=n;
    GotoXY(1,24); ClrEol;
  UNTIL NOT Yes('Once again? (Y/N) ');
END;

PROCEDURE FileResults(VAR f: TEXT);
VAR i: INTEGER;
BEGIN
   rewrite(f); Title(f);
   FOR i:=1 TO LastLetter do ShowLetterNo(f,i);
   close(f);
END;

VAR infile,outfile: TEXT;
BEGIN
  REPEAT ClearSums; ClrScr;
    GetAlphabet;  CapitalizeOrNot;
    WHILE GetFile(1,24,'Input text from file: ',infile) DO
    BEGIN Zero;
      CountThem(infile,LastLetter);
      FindVowels;
      Sort;
      Calculate;
      Title(output);
      ShowResults(LastLetter);
      IF GetFile(1,24,'Output results from '+nameOf(infile)+' to file: ',outfile)
      THEN FileResults(outfile);
    END;
    GotoXY(1,25);
  UNTIL NOT Yes('Another alphabet? (Y/N) ')
END.
----------------CUT HERE--------------------------

As you can see (loop WHILE GetFile etc...) you can "concatenate" files; really compute the statistics of a text contained into several files as if it were in a single file. It's not great software, but it could have been worse. An obvious improvement would be to allow for strings to be interpreted as single letters. Not difficult, but a problem to solve in conjunction with a BETTER user interface.

From gauss!G.GP.CS.CMU.EDU!Karl.Kluge Sun Dec  8 22:49 EST 1991
Received: by gauss; Sun Dec  8 23:54:50 EST 1991
Received: by inet.att.com; Sun Dec  8 23:54 EST 1991
Received: from G.GP.CS.CMU.EDU by rand.org; Sun, 8 Dec 91 20:29:56 -0800
Date: Sun, 8 Dec 1991 22:49-EST
From: Karl.Kluge@G.GP.CS.CMU.EDU
To: voynich@rand.org
Subject: Misc comments
Message-Id: <692250584/kck@G.GP.CS.CMU.EDU>
Status: OR

I think transcribing the figure captions is absolutely critical, in
particular to test hypotheses such as one of Brumbaugh's, which is that
the text is bogus but the captions are not. D'Imperio points out an
apparent difference in common word-initial characters, for instance.

Regarding choice of transcription alphabet: I think we have to be very
careful here, particularly with compound characters, as the assumptions
made in selecting the mapping may bias the results (I'm thinking in
particular of Tiltman's system considering K & F in the Currier
notation as the same char).

Merging J & 7? Best leave it alone in the already-transcribed text, and
hack code to fit if you think they're the same. Anyone know Currier and
can ask what the diff. is? (Jim G., you're mentioned in the NSA TR in a
way that implies you know D'Imperio -- can you ask her since she did
the transcription?)

Regarding the existing transcription -- looking at the Currier alphabet
in the back of D'Imperio, I don't see "#" (end of paragraph?), "$",
"*", or "," -- which Voynich chars are these? (I returned the TR to our
lib, so an explanation in terms of the alphabet in Guy's CRYPTOLOGIA
article would be best.)

Thanks to Guy for the explanation of Levitov's system. Just one minor
complaint -- you didn't give the correspondence between his characters
and either Currier's or your transcription alphabet, making it hard to
play along at home with the transcription. Also, when he says "polyglot
oral tongue," I think he has something like a creole in mind, which is
more specific than your suggestion and fits the sort of vocab and
grammar he gives.

Quick comment on Guy's CRYPTOLOGIA article on statistical properties,
specifically the attempt to identify vowels and consonants -- Tiltman
and D'Imperio seem to feel very strongly that the evidence is against
any sort of simple substitution cipher in an (unknown) natural
language, hence I'm not sure how trustworthy the vowel ident. results
are. Comments?

My pet theory is that it's a bilateral cypher with nulls. There are at
least four pairs of "dual" characters, P/F, H/K, Q/U, and V/Y; it may
explain the prefix/suffix structures found by Tiltman and the two
"hands" (stereotyped choices for bit pairs, differing between the two
scribes); it would also account for the low entropy (many Voynich chars
per text char). After that, I like D'Imperio's theory that it's an
artificial language along the lines of a papal code she describes. Then
comes Brumbaugh's theory as to the key to the captions, with the text
either bogus or further encoded.

The twenty most frequent words make me really suspicious (reverse
order): 8AE, 4OFAM, 4OFAE, ZC9, ZOE, 2, 8AN, SC9, OR, SOR, 4OFAN, 8AR,
4OFCC89, 4OFC89, 89, SOE, ZC89, OE, SC89, 8AM. For the 8th most common
word of a language to be 7 chars long strikes me as odd, at least based
on the 10 most common English words. Too much structure -- too many
4OF-'s, -89's, 8A-'s. Comments from the linguists on the list?

A bit of Voynich Mss. trivia -- in his "Return of the Lloigor" and THE
PHILOSOPHER'S STONE, Colin Wilson has characters who examine the Mss
and discover that it is the dreaded NECRONOMICON.

Sigh...back to thesis writing.

Karl

From gauss!G.GP.CS.CMU.EDU!Karl.Kluge Mon Dec  9 00:53 EST 1991
Received: by gauss; Mon Dec  9 02:31:52 EST 1991
Received: by inet.att.com; Mon Dec  9 02:31 EST 1991
Received: from G.GP.CS.CMU.EDU by rand.org; Sun, 8 Dec 91 22:56:57 -0800
Date: Mon, 9 Dec 1991 00:53-EST
From: Karl.Kluge@G.GP.CS.CMU.EDU
To: voynich@rand.org
Subject: Re: Misc comments
Message-Id: <692257998/kck@G.GP.CS.CMU.EDU>
Status: OR

> Reply-To: jim@rand.org
> Date: Sun, 08 Dec 91 21:16:04 PST
> 
> > Karl.Kluge@G.GP.CS.CMU.EDU writes:
> > I think transcribing the figure captions is absolutely critical, in
> 
> Agreed - Jim Reeds' suggestion looks eminently workable: marking locations
> with compass directions on the page, augmented with comments.

I can't find that mail message. (x, y) in mm relative to upper left
corner would be my preference, but suffers from problems of getting
copies with marked scale to everyone. Compass direction may be prob-
lematic for folios with many captions.

> > My pet theory is that it's a bilateral cypher with nulls. There are at
> > least four pairs of "dual" characters, P/F, H/K, Q/U, and V/Y; it may
> 
> Interesting -- like the Checkerboard cipher?

I mean like Francis Bacon's cipher using two type faces as (essentially)
binary 1 and 0, with a binary encoding of the alphabet -- so, for instance,
Voynich P, H, Q, and V might be 1's; F, K, U, and Y might be 0's; and
others might be nulls. D'Imperio discusses this type of code somewhere
in her TR. Kahn on supporters of the Bacon-as-Shakespeare theory also 
discusses it.

> > A bit of Voynich Mss. trivia -- in his "Return of the Lloigor" and THE
> > PHILOSOPHER'S STONE, Colin Wilson has characters who examine the Mss
> > and discover that it is the dreaded NECRONOMICON.
> 
> Neat -- John Dee wrote one of those... but it's gone missing from the UCLA
> Library, as do most things named Necronomicon from most libraries.

I suspect it hasn't gone missing, but was a joke. The MSU computer
catalog had an entry for the 1934 Randolph Carter edition of the Dee
translation, Randolph Carter being a character in a number of HPL's
stories. The Robert Turner NECRONOMICON pastiche, BTW, is a good read
for cryptographically oriented HPL fans.

However, Dee does lead into a theory that just occured to me re: the
origin of the Voynich MSS: It was Edward Kelly (in the drawing room 
with the candle stick).

Consider. One fine day, Kelly shows up on Dee's doorstep with two vials
of powder (capable of turning base metals to gold and silver respectively)
and a book "containing nothing but hieroglyphicks" that he found in a tomb 
(see D'Imperio for the details). Now, we can entertain two hypotheses here:
(1) some medieval Pons and Fleischman discovered "gold fusion" and Kelly 
happened upon their powder, or (2) Kelly was scamming Dee (see Chaucer for 
details of how this was usually done).

Later on, in Europe, Kelly (acting as crystal gazer for Dee) starts
dictating incantations in the so-called "Enochian" language, including
an Enochian script. Enochian has sufficient vocab and syntax for
Aleister Crowley to have translated the invocations from the GOETIA
into Enochian. Among other uplifting info from the spirits involved
were claims for the spiritual benefits to be reaped by Dee and Kelly
engaging in wife swapping. (D'Imperio also discusses Enochian, including
sample text.) Once again, we can entertain two theories: (1) Kelly really
was channelling spirits, or (2) Kelly made it all up.

If Kelly was capable of making up one bogus alphabet and language, why
not two? Has anyone computed the entropy of Enochian? My suspicion is
that it would turn out to have the same sort of anomolous low entropy
(on the theory that Kelly would have used a similar set of rules for
inventing the Voynich language and the Enochian language). One might
also examine Enochian for the sort of prefix/suffix structure Tiltman
found in the Voynich language. Unfortunately, if this theory is right,
there may be no cryptographic "smoking gun" to prove it (other than a
low entropy for Enochian, which would certainly be suggestive).

Brumbaugh has suggested Dee as the author in his book -- I think Dee 
was being duped by Kelly, who was delighted when Dee was able to get
600 ducats for what was originally nothing more than a prop.

Just my theory of the hour. 

Karl

From gauss!reeds Sun Dec  8 20:57:02 EST 1991
Received: by gauss; Sun Dec  8 21:09:08 EST 1991
Received: by inet.att.com; Sun Dec  8 21:09 EST 1991
Received: from research.att.com by rand.org; Sun, 8 Dec 91 17:57:14 -0800
Message-Id: <9112090157.AA22801@rand.org>
From: reeds@research.att.com
Date: Sun, 8 Dec 91 20:57:02 EST
To: voynich@rand.org
Status: OR

Hi, VMS gang!

My Newbold is at the office, so my version of N's plate II will have
to wait till the morning.

About alternate readings:  I had originally suggested using regular
expression syntax to denote alternate readings, thinking (but not
saying) regular expression in the sense of 'egrep'.  Thus, [abc] means
take exactly one of the single letters a b or c, and d? means take d
or not, and (efg|xyz) means take either the sequence efg or xyz; 
fancier formulas exist.  I think this covers what Jacques wants,
with a well established existing notation.

One trouble with such complicated formats is, that while they are 
expressive and precise, they are hard to use with naive programs.  
No longer will we be able to make a word frequency count with a 
single unix command like 

sed -e 's/^.......//' -e 's/;.*//'< data|tr '-/ ' \12|sort|uniq -c|sort -rn

say, nor will we be able to write little 10 line digraph counting routines
the dumb way, whose business end is the sequence

	while((c=getchar())!=EOF) { paircount[oc][c]++; oc = c; }

and so on and so forth.  Experimentation will become somewhat more expensive,
and our programs will have to access the data through special purpose access
routines.  Which we will have to write.

Even the job of automatically checking to see if two transcriptions of 
the same text are equivalent becomes harder.  If Jacques transcribes it 
as (a|bc)(de|f) and Gilles as (bcde|af|ade|bcf), it takes a moment or
two to puzzle out the answer.  Its possible to do the comparison auto-
matically, but the program to do it is not trivial.

Line numbers and other locus information ARE comments, as Jacques suggests,
but not arbitrary comments.  It is legal to have spelling mistakes in ordinary
comments, say, but not in line numbers: they must be as precisely parseable as 
the representation of the text itself.

I have partial sympathy for the idea of bracketed comments, but the 
same objections apply.  They drive me crazy in programming languages,
though, when put inside expressions, as in 

	sin(/*theta already in radians*/theta+PI*/*but a is not*/a/180)
	
I suppose experimentation with all proposals is the way to tell.

Happy St. Oladabas day!

From gauss!rand.org!jim%mycroft Sun Dec 08 18:07:14 PST 1991
Received: by gauss; Sun Dec  8 21:12:06 EST 1991
Received: by inet.att.com; Sun Dec  8 21:12 EST 1991
Received: from mycroft.rand.org by rand.org; Sun, 8 Dec 91 18:07:18 -0800
Received: from localhost by mycroft.rand.org; Sun, 8 Dec 91 18:07:15 PST
Message-Id: <9112090207.AA03787@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: ground rules proposal: comments and page numbers
In-Reply-To: Your message of Mon, 09 Dec 91 10:51:06 -0500.
             <9112082351.AA08729@medici.trl.OZ.AU> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Sun, 08 Dec 91 18:07:14 PST
Sender: jim%mycroft@rand.org
Status: OR

Jacques Guy suggests that we need bracketed comments also; his arguments
make sense to me.  I'd prefer one flavor, rather than both {} and ;, so
I vote for the {} braces.

Agree also about the | separator, as in (sort of) BNF, and with ' ' as
a no-op.

I don't agree that the page numbers should be in the same kind of comment
brackets, since they include substantive information about the location;
this includes Jim Reeds' proposal about directional identification on the
picture captions -- the locations on the pages might well be meaningful,
and a program should be able to count on it in a non-comment field.

However, they aren't really part of the text -- how about using the angle
brackets <> to set off the page/line/location information?  What should we
do about the "hand" and "language"?  That's sort of a second-order piece
of information.

I'm not too wild about the probability estimates on ambiguous characters,
but am willing to be convinced if others think it makes sense.

That would make my 65v look like this:

{ Folio 65v (Herbal) appears as Plate II in Newbold.
  It features a plant with many roots, branches and flowers; the main
  stem separates the lines of text, which are well-registered across
  the stem.
  Transcribed 7 Dec 91 by Jim Gillogly.
}
<065V01> W9/ VSCY9/ 89/-/ 8SCBAN/ ZCP9/ 4OB9/ VOE/ SB89-
<065V02> 8AM/ ZCCF/ E/ O89/-/ 9PCO/ 4OB[*|R]/ AT/ SCOPCC9/ 8AEAJ-
<065V03> PAE/ SCOP/ ZF9/ 9/-/ [*|S|R|2]AR/  SCO89-
<065V04> POCC89/ OPOR/ ZC89/-/ 9PCC9/ ZCO89/ 4OFZ8/ 9VSC9-
<065V05> 8ZC89/ OFCO89/ 4OF8/-/ ZX9/ SOF9/ SOFCO89/ OFC9/ 89-
<065V06> [4|9]O/ 9PSC9/ ZXCW9/-/ 9FSC89/ SC89/ ZX89#


Jim Gillogly

From gauss!reeds Sun Dec  8 21:11:34 EST 1991
Received: by gauss; Sun Dec  8 21:51:34 EST 1991
Received: by inet.att.com; Sun Dec  8 21:51 EST 1991
Received: from research.att.com by rand.org; Sun, 8 Dec 91 18:11:47 -0800
Message-Id: <9112090211.AA22962@rand.org>
From: reeds@research.att.com
Date: Sun, 8 Dec 91 21:11:34 EST
To: voynich@rand.org
Status: OR

Yet more on variant readings, space vs slash, & so on.  Why not 
use "." to represent word space (its a lighter character than "/")
and use "," as an uncertain word space, which might otherwise be 
represented as /? or (|/) in regular expression notation.   If you
call "." a "full stop" you can indoctrinate yourself into 
calling "," a "half stop", which gives the current proposal some
color of rationality.

I suspect that most variant readings will hinge on the uncertain 
word space.

From gauss!rand.org!jim%mycroft Sun Dec 08 18:18:47 PST 1991
Received: by gauss; Sun Dec  8 21:53:58 EST 1991
Received: by inet.att.com; Sun Dec  8 21:53 EST 1991
Received: from mycroft.rand.org by rand.org; Sun, 8 Dec 91 18:18:51 -0800
Received: from localhost by mycroft.rand.org; Sun, 8 Dec 91 18:18:48 PST
Message-Id: <9112090218.AA03833@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: ground rules proposal: resources
In-Reply-To: Your message of Mon, 09 Dec 91 10:51:06 -0500.
             <9112082351.AA08729@medici.trl.OZ.AU> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Sun, 08 Dec 91 18:18:47 PST
Sender: jim%mycroft@rand.org
Status: OR

Yes, it would make sense to know what kind of equipment we have.  I suspect
that for the most part we'll be writing our own hacks, but here's what I've
got:

	PC clone (386) with VGA, Epson dot matrix printer
	Sun Sparcstation 1, Postscript printers and HP laserjet sometimes
	Occasional Mac II access

	Copy of BM microfilm, which includes ff 1r-11v, 13r-55v
	Newbold
	Kahn
	D'Imperio
	Brumbaugh
	various unpublished stuff, newspaper clippings, journal articles

I'm fluent in C and Perl, adequate in Turbo Pascal, and have a vast store
of crypto programs for handling standard hobby stuff (Playfair, Vigenere,
Bifid, and like that), few of which are likely to be relevant here.

I don't have any experience with a mixed text/graphics package that
would be useful here.  Anybody?

I'd like to see a Voynich font for the PC (EGA or better) and Epson
printer if anybody's into that...


Jim Gillogly

From gauss!rand.org!jim%mycroft Sun Dec 08 18:33:03 PST 1991
Received: by gauss; Sun Dec  8 21:56:15 EST 1991
Received: by inet.att.com; Sun Dec  8 21:56 EST 1991
Received: from mycroft.rand.org by rand.org; Sun, 8 Dec 91 18:33:06 -0800
Received: from localhost by mycroft.rand.org; Sun, 8 Dec 91 18:33:04 PST
Message-Id: <9112090233.AA03888@mycroft.rand.org>
Subject: Variant readings
To: voynich@rand.org
In-Reply-To: Your message of Sun, 08 Dec 91 20:57:02 -0500.
             <9112090157.AA22801@rand.org> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Sun, 08 Dec 91 18:33:03 PST
Sender: jim%mycroft@rand.org
Status: OR


> reeds@research.att.com writes:
> 
> About alternate readings:  I had originally suggested using regular
> expression syntax to denote alternate readings, thinking (but not
  ...
> One trouble with such complicated formats is, that while they are 
> expressive and precise, they are hard to use with naive programs.  

Yeah, that would make hacking pretty interesting, all right.  I'm not
sure but what general regular expressions are overkill for the kind of
ambiguities we expect.  Do you see much text that would require one
of your horror-story expressions, rather than the simpler ones suggested
by Jacques?

For simple-minded programs we could write a quick Perl hack to turn it all
into the least-common-denominator format, using the order of options to
select the most likely (new convention: put the most likely first).  That
would be easy to write (if we're not talking full re's), and Perl is
available both on Unix and PC.

> Even the job of automatically checking to see if two transcriptions of 
> the same text are equivalent becomes harder.  If Jacques transcribes it 

Hmm.  That's a lose, isn't it.

> Happy St. Oladabas day!

And a joyous Michiton to you as well.  (Is St. Dunstan's day too obscure?)

Jim Gillogly

From gauss!rand.org!jim%mycroft Sun Dec 08 20:42:35 PST 1991
Received: by gauss; Sun Dec  8 23:48:56 EST 1991
Received: by inet.att.com; Sun Dec  8 23:48 EST 1991
Received: from mycroft.rand.org by rand.org; Sun, 8 Dec 91 20:42:38 -0800
Received: from localhost by mycroft.rand.org; Sun, 8 Dec 91 20:42:35 PST
Message-Id: <9112090442.AA04072@mycroft.rand.org>
To: voynich@rand.org
Subject: Voynich languages (dialects?)
In-Reply-To: Your message of Mon, 09 Dec 91 15:17:24 -0500.
             <9112090417.AA08963@medici.trl.OZ.AU> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Sun, 08 Dec 91 20:42:35 PST
Sender: jim%mycroft@rand.org
Status: OR

> j.guy@trl.OZ.AU (Jacques Guy) writes:
> Herbal with 2 color reproductions from Voynich

What's this one?  Is it generally available?  Does that mean the whole
Herbal section?

We're talking a bit at cross-purposes on the "language" issue.  So far as
I know there's not much variation in the externally-visible language:
mostly Voynich Alphabet, with a note or two in German, the Latin zodiac
names, and the mixed German(?)/ Voynich/Latinoid "key" -- perhaps not
enough to worry about detailed conventions.

I was referring to Currier's observation that different "language" was
used in different sections -- I posted a paragraph of his a couple days
ago referring to a different "Pharmaceutical B" vs "Biological B": they
are written in the same "hand", but use different "language", according
to his statistics.  This is the kind of thing I was thinking of as a
second-order effect.  It might be considered a different dialect; or
perhaps simply reflects different vocabulary for different subject matter.
I think we could just pop this in as a normal undistinguished comment.

OK on the '.' and ',' for spaces and possible spaces.


Jim Gillogly

From gauss!rand.org!jim%mycroft Sun Dec 08 21:16:04 PST 1991
Received: by gauss; Mon Dec  9 00:22:30 EST 1991
Received: by inet.att.com; Mon Dec  9 00:22 EST 1991
Received: from mycroft.rand.org by rand.org; Sun, 8 Dec 91 21:16:08 -0800
Received: from localhost by mycroft.rand.org; Sun, 8 Dec 91 21:16:05 PST
Message-Id: <9112090516.AA04158@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: Misc comments 
In-Reply-To: Your message of Sun, 08 Dec 91 22:49:00 -0500.
             <692250584/kck@G.GP.CS.CMU.EDU> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Sun, 08 Dec 91 21:16:04 PST
Sender: jim%mycroft@rand.org
Status: OR


> Karl.Kluge@G.GP.CS.CMU.EDU writes:
> I think transcribing the figure captions is absolutely critical, in

Agreed - Jim Reeds' suggestion looks eminently workable: marking locations
with compass directions on the page, augmented with comments.

> hack code to fit if you think they're the same. Anyone know Currier and
> can ask what the diff. is? (Jim G., you're mentioned in the NSA TR in a
> way that implies you know D'Imperio -- can you ask her since she did
> the transcription?)

I'll send a letter to the last address I have for her -- haven't talked
to her in 10 years, though.  I met her while I was on temporary assignment
at RAND's Washington office in the mid-70s.

> Regarding the existing transcription -- looking at the Currier alphabet
> in the back of D'Imperio, I don't see "#" (end of paragraph?), "$",
> "*", or "," -- which Voynich chars are these? (I returned the TR to our
> lib, so an explanation in terms of the alphabet in Guy's CRYPTOLOGIA
> article would be best.)

There's a file in the ftp directory called "currier" that discusses these
issues.  Briefly, '-' is a line break or a broken line, '#' is a paragraph
break, '*' appears to be illegible, and ',' appears to be an error -- in
at least one case it should have been an 8, which is on the same keypunch
key.

> language, hence I'm not sure how trustworthy the vowel ident. results
> are. Comments?

If there's a credible vowel/consonant split, it could at least give us
a pronounceable way to read the text... for mnemonic convenience.

> My pet theory is that it's a bilateral cypher with nulls. There are at
> least four pairs of "dual" characters, P/F, H/K, Q/U, and V/Y; it may

Interesting -- like the Checkerboard cipher?

> The twenty most frequent words make me really suspicious (reverse
> order): 8AE, 4OFAM, 4OFAE, ZC9, ZOE, 2, 8AN, SC9, OR, SOR, 4OFAN, 8AR,
> 4OFCC89, 4OFC89, 89, SOE, ZC89, OE, SC89, 8AM. For the 8th most common
> word of a language to be 7 chars long strikes me as odd, at least based
> on the 10 most common English words. Too much structure -- too many
> 4OF-'s, -89's, 8A-'s. Comments from the linguists on the list?

I'm not a linguist, but it doesn't look weird to me.  For one thing, the
40, the CC, and the 89 are common enough that they may all be single
characters.  For another, the 25th most common word in the "digest" for
this mailing list so far is "Voynich" and the 32nd most common is
"transcription".  Maybe that word is central to the subject matter.

> A bit of Voynich Mss. trivia -- in his "Return of the Lloigor" and THE
> PHILOSOPHER'S STONE, Colin Wilson has characters who examine the Mss
> and discover that it is the dreaded NECRONOMICON.

Neat -- John Dee wrote one of those... but it's gone missing from the UCLA
Library, as do most things named Necronomicon from most libraries.  Part
of the cosmic unconsciousness, as Miller observed in Repo Man, I suspect.

Jim Gillogly

From gauss!reeds Mon Dec  9 00:25:59 EST 1991
Received: by gauss; Mon Dec  9 00:28:52 EST 1991
Received: by inet.att.com; Mon Dec  9 00:28 EST 1991
Received: from research.att.com by rand.org; Sun, 8 Dec 91 21:26:09 -0800
Message-Id: <9112090526.AA26324@rand.org>
From: reeds@research.att.com
Date: Mon, 9 Dec 91 00:25:59 EST
To: voynich@rand.org
Status: OR

My equipment:  departmental computers at work, mostly big VAXes, SGI
machines, a Cray in a pinch.  B&W laser printers.  I am fluent in C
and FORTRAN; am a complete UNIX technogeek.  I have a lot of smart
colleagues I can ask for help, if I need to.

My sources:  a bunch of reprints, D'Imperio, the usual crypto books.
I intend to order a film of the MS.

Karl Kluge asks

> Regarding the existing transcription -- looking at the Currier alphabet
> in the back of D'Imperio, I don't see "#" (end of paragraph?), "$",
> "*", or "," -- which Voynich chars are these? 

The transcription seems to have lots of errors, maybe 1 char in 200
is wrong. $ and , seem to be keypunching errors.  * means illegible,
and # is sometimes used as a paragraph indicator.


Jim Gillogly asks

>  
>  > j.guy@trl.OZ.AU (Jacques Guy) writes:
>  > Herbal with 2 color reproductions from Voynich
>  
>  What's this one?  Is it generally available?  Does that mean the whole
>  Herbal section?

Presumably Wilfrid Blunt & Sandra Raphael, ``The Illustrated Herbal'',
ISBN 0-500-1226-1, Thames and Hudson, 1979.  On p 88 are indistinct
(black on brown!) shots of 78v, 79r, 99v and 100r, and on pp90-91
is a superb color 33v and 34r.

Jim Reeds.

From gauss!gauss.att.com!reeds Mon Dec  9 02:01:09 EST 1991
Received: by gauss; Mon Dec  9 02:33:25 EST 1991
Received: by inet.att.com; Mon Dec  9 02:33 EST 1991
Received: from research.att.com by rand.org; Sun, 8 Dec 91 23:01:23 -0800
Message-Id: <9112090701.AA28066@rand.org>
From: reeds@gauss.att.com
Date: Mon, 9 Dec 91 02:01:09 EST
To: voynich@rand.org
Status: OR

Here is an interim version of my list of Voynich pages.
I have a bunch of extra image locations to fill in, and want
to cross check some stuff from Newbold.  And of course I
have got to fill out the bibliographical data for the
citations.

But in the mean time you can use this check list.

If you know of any more published V images, or see any
mistakes here, please let me know.

DRAFT!!!

Voynich MS foliation

Format: column1		 column2	; column3
	
	column 1: page numbers according to D'Imperio's book

	column 2: location:
		plain numbers = page numbers in D'Imperio's transcription
		BL = British Library photocopies (Gillogly has a copy)

	column 3: ; description

		A  = Currier's hand A (from computer file)
		B  = Currier's hand B (from computer file)

		Notes from from D'Imperio, fig.4, p.82:
		h  =  has ``herbal'' pictures
		c  =  has ``cosmological'' pictures
		a  =  has ``astronomical'' pictures
		p  =  has ``pharmaceutical'' pictures
		m  =  has figures of humans
		T  =  text only
		K  =  has ``key like'' sequences
		E  =  has some writing in extraneous (non V) script

Goal:
	A unique unambiguous name for each page in the VMS,
	to unobtrusively mesh with the existing foliation
	on the MS.

Problems:
	1. Kraus's "35 MSS" lists 7 double folding folios,
	  3 triples, and 1 quadruple.  Which are they?
	  (Context in D'Imperio indicates that ff 3, 72, and 73
	  are the triples, that f 86 is the quadruple, and
	  that ff 67, 70, 89, 90, 101, and 102 are doubles.
	  This leaves one double unaccounted for.)

	2. Kraus lists ff 59-64 as missing, Newbold does not.
	  (Did they get lost in the intervening half-century?)

	3. The D'Imperio transcription lists pages serially
	  between 1-111 and 147-166.  The gap 112-146 is
	  2 pages off the calculated size.  Why?  (75r should
	  seemingly be p149 not p147.)

	4. D'Imperio refers to loci like 85/86r2, presumably
	  meaning the 2nd part of the quadruple foldout, recto.
	  Which folio is the quad, 85 or 86?  How is the
	  folding done: accordion style or into quarters;
	  how do the parts get numbered?  What's worse, she
	  refers to 85/86v5 and  85/86v6, which is odd, I say,
	  very odd.

References to image and transcription loci:

	Bennett
	Blunt & Raphael, The Illustrated Herbal
	BL film (not seen)
	Brumbaugh, Speculum 49, 1974, p546-8. Botany and the Voynich
	``Roger Bacon'' manuscript once more.
	Brumbaugh, The Yale University Library Gazette, p347-335..
	The solution of the Voynich ``Roger Bacon'' cipher.
	Brumbaugh, book, not seen
	Brumbaugh, J Courtald, not seen
	D'Imperio, computer transcription
	Guy, Cryptologia
	Kahn, The Codebreakers
	Manly, Speculum
	Newbold
	Yale film (not seen)
	Zimansky (not seen)

Acknowledgements and sources:
	
	D'Imperio, The Voynich Manuscript - an elegant enigma
	
	inspection of the transcription file at rand
	
	personal communication from Gillogly

	Thanks, and a tip of the hat to Mary and Jim!


Jim Reeds
7 Dec 1991


1r      1 BL			; A T K E
1v      2 BL			; A h
2r      3 BL			; A h
2v      4 BL			; A h
3r      5 BL			; A h
3v      6 BL			; A h
4r      7 BL			; A h
4v      8 BL			; A h
5r      9 BL			; A h
5v      10 BL			; A h
6r      11 BL			; A h
6v      12 BL			; A h
7r      13 BL			; A h
7v      14 BL			; A h
8r      15 BL			; A h
8v      16 BL			; A h
9r      17 BL			; A h
9v      18 BL			; A h
10r     19 BL			; A h
10v	20 BL			; A h
11r     21 BL			; A h
11v     22 BL			; A h
12 missing
13r     23 BL			; A h
13v     24 BL			; A h
14r     25 BL			; A h
14v     26 BL			; A h
15r     27 BL			; A h
15v     28 BL			; A h
16r     29 BL			; A h
16v     30 BL			; A h
17r     31 BL			; A h E
17v     32 BL			; A h
18r     33 BL			; A h
18v     34 BL			; A h
19r     35 BL			; A h
19v     36 BL			; A h
20r     37 BL			; A h
20v     38 BL			; A h
21r     39 BL			; A h
21v     40 BL			; A h
22r     41 BL			; A h
22v     42 BL			; A h
23r     43 BL			; A h
23v     44 BL			; A h
24r     45 BL			; A h
24v     46 BL			; A h
25r     47 BL			; A h
25v     48 BL			; A h
26r     49 BL			; A h
26v     50 BL			; A h
27r     51 BL			; A h
27v     52 BL			; A h
28r     53 BL			; A h
28v     54 BL			; A h
29r     55 BL			; A h
29v     56 BL			; A h
30r     57 BL			; A h
30v     58 BL			; A h
31r     59 BL			; A h
31v     60 BL			; A h
32r     61 BL			; A h
32v     62 BL			; A h
33r     63 BL			; A h
33v	64 BL NYT, 6 May 75 (Brumbaugh) Kahn plate, near p.845, B&R p90 ; B h
34r     65 BL B&R p91		; B h
34v     66 BL			; B h
35r     67 BL			; A h
35v     68 BL			; A h
36r     69 BL			; A h
36v     70 BL			; A h
37r     71 BL			; A h
37v     72 BL			; A h
38r     73 BL			; A h
38v     74 BL			; A h
39r     75 BL			; B h
39v     76 BL			; B h
40r     77 BL			; B h
40v	78 BL Child Sample Readings, p.17 ; B h
41r     79 BL			; B h
41v     80 BL			; B h
42r     81 BL			; A h
42v     82 BL			; A h
43r     83 BL			; B h
43v     84 BL			; B h
44r     85 BL			; A h
44v     86 BL			; A h
45r     87 BL			; A h
45v     88 BL			; A h
46r     89 BL			; B h
46v     90 BL			; B h
47r     91 BL			; A h
47v     92 BL			; A h
48r     93 BL			; B h
48v     94 BL			; B h
49r	95 BL Brumbaugh p.4, Harper's Jul 1921, 186-97 ; A h K
49v     96 BL			; A h
50r     97 BL			; B h
50v     98 BL			; B h
51r     99 BL			; A h
51v     100 BL			; A h
52r     101 BL			; A h
52v     102 BL			; A h
53r     103 BL			; A h
53v     104 BL			; A h
54r     105 BL			; A h
54v     106 BL			; A h
55r     107 BL			; B h
55v     108 BL			; B h
56r     109 BL			; A h
56v     110 BL			; A h
57r     111 BL			; B h
57v				; c K E
58r				; T
58v				; T
59 missing
60 missing
61 missing
62 missing
63 missing
64 missing
65r					; h
65v	Newbold plate I			; h
66r	Brumbaugh Yale Lib Gaz p352	; T K E
66v					; h
67r1	Newbold plate III		; a
67r2					; a
67v1					; a
67v2					; c
68r1					; a
68r2					; a
68r3					; a
68v1					; a
68v2					; a
68v3					; c
69r					; c
69v					; c
70r1					; c
70r2					; Pisces
70v1					; Aries (dark)
70v2
71r					; Aries (light)
71v	Taurus (light) Newbold plate XXIV (mislabeled 79v) ; Turus (light)
72r1					; Taurus (dark)
72r2					; Gemini
72r3					; Cancer
72v1					; Libra
72v2					; Virgo
72v3	Newbold plate XXV		; Leo 
73r1
73r2
73r3					; Scorpio
73v1
73v2
73v3					; Sagittarius
74 missing				; Capricorn, Aquarius?
75r     147				; B m
75v     148				; B m
76r	149 reversed in Brumbaugh p.114 ; B T K
76v     150				; B m
77r     151				; B m
77v     152 Kahn p86			; B m
78r     153 Newbold plate V		; B m
78v     154 B&R p88			; B m
79r     155 B&R p88			; B m
79v	156 Guy p.212 after Bennett p.188 (mislabelled in Bennett) ; B m
80r	157 Guy p.213 after Bennett p.189 (mislabelled in Guy) ; B m
80v	158				; B m
81r	159 Brumbaugh p.9, Harper's	; B m
81v     160				; B m
82r     161				; B m
82v	162 Brumbaugh p.104		; B m
83r     163				; B m
83v     164				; B m
84r     165 Newbold VI			; B m
84v     166				; B m
85/86r1					; T
85/86r2					; c
85/86r3					; r
85/86r4					; r
85/86v1	Newbold plate VII, VIII		; r
85/86v2					; r
85/86v3					; c
85/86v4					; c
85/86v5					; T
85/86v6					; T
85r	Newbold plate VIII, Brumbaugh p.54
85v	Newbold plate VIII
86v	Newbold plate VII, VIII
87r					; h
87v					; h
88r Washington Post, 5 Aug 62 (Sunday)	; p
88v					; p
89r1					; p
89v1					; p
89r2					; p
89v2					; p
90r1					; h
90v1					; h
90r2					; h
90v2					; h
91 missing
92 missing
93r		Brumbaugh p80		; h
93v					; h
94r					; h
94v					; h
95r					; h
95v					; h
96r					; h
96v					; h
97 missing
98 missing
99r					; p
99v	B&R p88				; p
100r	B&R p88, Brumbaugh Speculum p.547, Brumbaugh p.96	;p
100v	Brumbaugh p.86			; p
101r					; p
101r1					; p
101r2					; p
101v1					; p
101v2					; p
102r1					; p
102r2					; p
102v1					; p
102v2					; p
103r					; T stars
103v					; T stars
104r					; T stars
104v					; T stars
105r					; T stars
105v					; T stars
106r					; T stars
106v					; T stars
107r					; T stars
107v					; T stars
108r					; T stars
108v					; T stars
109 missing
110 missing
111r					; T stars
111v					; T stars
112r					; T stars
112v					; T stars
113r					; T stars
113v					; T stars
114r					; T stars
114v					; T stars
115r					; T stars
115v					; T stars
116r					; T stars
116v					; K E

----- end of document -----

From gauss!rand.org!jim%mycroft Mon Dec 09 04:51:29 PST 1991
Received: by gauss; Mon Dec  9 08:28:59 EST 1991
Received: by inet.att.com; Mon Dec  9 08:28 EST 1991
Received: from mycroft.rand.org by rand.org; Mon, 9 Dec 91 04:51:36 -0800
Received: from localhost by mycroft.rand.org; Mon, 9 Dec 91 04:51:30 PST
Message-Id: <9112091251.AA04589@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: Misc comments 
In-Reply-To: Your message of Mon, 09 Dec 91 00:53:00 -0500.
             <692257998/kck@G.GP.CS.CMU.EDU> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Mon, 09 Dec 91 04:51:29 PST
Sender: jim%mycroft@rand.org
Status: OR


> Karl.Kluge@G.GP.CS.CMU.EDU writes:
> However, Dee does lead into a theory that just occured to me re: the
> origin of the Voynich MSS: It was Edward Kelly (in the drawing room 
> with the candle stick).

That's been my theory for quite a few years now.  I was poking around at it
in the UCLA library a few weeks ago -- turns out there is some fairly new
work on Enochian, making it more accessible than going to the Sloane
Manuscripts in the BM.

The book I was studying is

	John Dee's Actions With Spirits
		25 December 1581 to 23 May 1583
	Christopher Whitby
	Garland Publishing Co., NY & London 1988
	ISBN 0-8240-6399-6

It's got a lot of detail on Enochian, and includes Kelly's discovery of
the cipher manuscript -- they occasionally called it the Book of St.
Dunstan, or something to that effect, which is the source of my somewhat
oblique remark about St. Dunstan the other day.

The book (also referred to as a scroll) was brought to Dee by Kelly and a
Mr. Husy or Husey in 1583, and was the subject of a couple of discussions
between Kelly/Dee and the angels... who weren't forthcoming about its
meaning.

One item on my agenda is to write to the BM for a couple of prints from
the Sloane Ms collection that shows Kelly's and Dee's handwriting... it's
my conjecture that one of the "hands" is Kelly's, and the other is some
unnamed co-conspirator (possibly Husy?).

Even if it *is* Kelly's fabrication, that doesn't mean it's meaningless.
Enochian is his work (apologies to those who are into angels), and is
said to hang together very well by the guy who did the Enochian dictionary
(also (?) gone missing from the UCLA collection, and I can't find the
reference right now).

> not two? Has anyone computed the entropy of Enochian? My suspicion is
> that it would turn out to have the same sort of anomolous low entropy
> (on the theory that Kelly would have used a similar set of rules for
> inventing the Voynich language and the Enochian language). One might

I've accumulated a fair bunch of Enochian, but haven't typed it in yet.
Here's a sample, from page 16 of the Book of Enoch (I think), Sloane 3188
f 74a:

	Gesco a taffom ges nat gam  pamphe ordaquaf
	cesto kidmap mischna iaiag  iaialpazudph a
	damset vnban caf ransembloh   dafna vp
	aschem gras chramsa asco dah  viana
	gen alde os papeam och lauan vnad.
	Oh drosad udrios nagel panzo ab sescu.
	Vorge afcal valaffda mormab gaf ham de
	Peleh asca.

I don't guarantee this stuff -- the copy I'm reading from has the m's and n's
blurred, for example, and I can't really tell the difference.  I doubt that
there's a connection between the "gaf" in the penultimate line and the one
in the Voynich "key", but who knows...

Again, I haven't run stats on it, but this stuff looks like it has higher
entropy than Voynich Language.  That's on my agenda too, though.


Jim Gillogly

From gauss!gauss.att.com!reeds Mon Dec  9 09:42:21 EST 1991
Received: by gauss; Mon Dec  9 09:53:58 EST 1991
Received: by inet.att.com; Mon Dec  9 09:53 EST 1991
Received: from research.att.com by rand.org; Mon, 9 Dec 91 06:42:37 -0800
Message-Id: <9112091442.AA05831@rand.org>
From: reeds@gauss.att.com
Date: Mon, 9 Dec 91 09:42:21 EST
To: voynich@rand.org
Status: OR

Here is my attempt.  

<65v>{Newbold II facing p44}
{Reeds 9 dec 91}
{Plant divides text into 2 columns, with apparant line continuation}

<65v.1>(W9.VSCW9.89)	  {stem} (8SBAN.ZCP9.4OB9.VOE.SB89) {ghost 4 under 4 of 4OB9}
<65v.2>(8AM.ZCCF,E,O89)   {stem} (9PCO.4OB2.AT.SCOPCCO.8AEOJ)
<65v.3>(9PAE.SCOPZF9.9)   {stem} (2A2.SCO89) {initial 2 of 2A2 doubtful}
{
para break on both sides of plant
}
<65v.4>(POCC89.OPOR,ZC89) {stem} (9PCC9.ZCO89.4OFZ8.9VSC9)
<65v.5>(8ZC89.OFCO89.4OF8){stem} (ZX9.SOF9.SOFCO89.OFC989)
<65v.6>(4O.9PSC9.ZXCW9)   {stem} (9FSC89.SC89.ZCFC89)

From gauss!math.mit.edu!jbaez Mon Dec  9 10:18:37 EST 1991
Received: by gauss; Mon Dec  9 10:35:36 EST 1991
Received: by inet.att.com; Mon Dec  9 10:35 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Mon, 9 Dec 91 07:21:19 -0800
Received: from cayley (CAYLEY.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA14041; Mon, 9 Dec 91 10:17:24 EST
Received: by cayley; Mon, 9 Dec 91 10:18:37 EST
Date: Mon, 9 Dec 91 10:18:37 EST
From: jbaez@math.mit.edu
Message-Id: <9112091518.AA19514@cayley>
To: voynich@rand.org
Subject: Introduction, etc.
Status: OR

Thanks for the introductions.  Again, I'm John Baez, and I'm a
mathematical physicist at U. C. Riverside, visiting Wellesley College
this year, and hanging out at MIT much of the time.  If quantum field
theory turns out to be the key to the Voynich manuscript I will be a real
asset to this group.  I can just barely program my way out of paper bag
in number of languages, but I have recruited Nate Osgood, a computer science
grad student at MIT, who may make up for that deficiency.  Together we
have access to plenty of hardware: Sparcstations and Suns at the math dept.
of MIT, a scanner in LCS (Laboratory for Computer Science), etc., and I 
have a Mac II in my office and PC 386 clone at home.

About transcription systems I have some general remarks:

1) it should be easy to type masses of stuff in quickly, and without 
requiring more than any old text processor.  For example, when I'm 
hanging out at my fiancee's place in New York I should be able to 
transcribe away on her laptop.   GIF-based systems etc. would rule this out.

2) it should be easy to read.  I find Prescott Currier's system excellent
in this respect: one can look at his transcription and the ms and easily
mentally check it.  While computers promise to be the reason why we succeed
where so many others failed (?), having a feel for the ms is crucial too.
Transcriptions `a la grep' make things harder on humans.

Am I right in thinking that only Jim actually has photocopies of the
whole ms right now, the British Library version?  I am in the process of
trying to get the Yale version.  It seems crucial for us all to quickly get
identical sets of one or the other.  (It'd be great if someone could get
Aegean Park Press to publish it!  Does anyone know them well?  Would Yale or
the British Library scream?)  Once we have such standardized access to the
real thing, high-tech solutions involving GIF's etc. would seem unnecessary.

(That is, given the difficulties.)

3) I don't mind periods too much for indicating spaces.  Is there a good
reason for not using spaces?  

John Baez

From gauss!rand.org!jim%mycroft Mon Dec 09 07:40:59 PST 1991
Received: by gauss; Mon Dec  9 11:11:39 EST 1991
Received: by inet.att.com; Mon Dec  9 11:11 EST 1991
Received: from mycroft.rand.org by rand.org; Mon, 9 Dec 91 07:41:03 -0800
Received: from localhost by mycroft.rand.org; Mon, 9 Dec 91 07:41:00 PST
Message-Id: <9112091541.AA04986@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: Introduction, etc. 
In-Reply-To: Your message of Mon, 09 Dec 91 10:18:37 -0500.
             <9112091518.AA19514@cayley> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Mon, 09 Dec 91 07:40:59 PST
Sender: jim%mycroft@rand.org
Status: OR


> jbaez@math.mit.edu writes:
> Am I right in thinking that only Jim actually has photocopies of the
> whole ms right now, the British Library version?  I am in the process of

Nope.  I requested a copy from the BM, and they were kind enough to send
it.  The microfilm I received is ff 1r-55r only (I mistakenly put 55v in
my previous message), except for the missing f 12.  I didn't write back to
ask whether that was all of it, but no reason to believe they wouldn't
have sent the other 2/3.  All of these are in the D'Imperio transcription,
so their main early use will be to check that part of the work.  And yes,
our transcription won't get far if we can't get the pictures to
transcribe!

Let us know when the Beinecke writes back -- I'm sure we'd all like to know
how much it will cost and how good the copies are.

> the British Library scream?)  Once we have such standardized access to the
> real thing, high-tech solutions involving GIF's etc. would seem unnecessary.

I disagree.  If we can afford the size of GIFs they will make it a lot easier
for us to communicate about specific parts -- clip a woman in a vat with the
caption, rather than trying to describe where in our jointly-shared
hard-copy archive it is.  Also, note that there are now 15 people on the
list.  Distributing hard-copy stuff is going to get prohibitive very fast!

> 3) I don't mind periods too much for indicating spaces.  Is there a good
> reason for not usingspaces?  

One good one -- you can spread the text out with program-ignored whitespace
to do your inter-linear commenting and translating.  The distinction between
certain and possible spaces is also worth maintaining.

[This posting deserves a fair number of IMHOs, I think... consider it read.]

Jim Gillogly

From gauss!trl.OZ.AU!j.guy Mon Dec  9 10:51:06 EST 1991
Received: by gauss; Sun Dec  8 19:09:10 EST 1991
Received: by inet.att.com; Sun Dec  8 19:09 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sun, 8 Dec 91 15:52:30 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA08729; Mon, 9 Dec 91 10:51:06 EST
Date: Mon, 9 Dec 91 10:51:06 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112082351.AA08729@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Re: ground rules proposal
Status: OR


Easy things first.

 Everybody agree with "Semicolon initiates comment to end of line"?

 *I* do NOT! I know, I know, it was what I used in my little vowelfq program, but
that's no reference. We need bracketed comments too, that can appear anywhere
in the text. I suggest either {} or <>. References to folios and lines ought
to be bracketed in that manner. Two good reasons: first conceptually, they ARE 
comments, and second, it makes for programs much, much easier to write.

Alternative readings. We need a separator. | is the perfect candidate. Why a separator?
Precisely because we do not know for sure where letters start and end, especially
where the script is lightly connected. For instance, there will be passages where
we are not sure whether we have CM or AN (Currier's system). So: [CM|AN]. 
Without a separator, it's awfully ambiguous.

Here is, as an example, once again Jim Gillogly's transcription of 65v as it stands:

; Folio 65v (Herbal) appears as Plate II in Newbold.
; It features a plant with many roots, branches and flowers; the main
; stem separates the lines of text, which are well-registered across
; the stem.
; Transcribed 7 Dec 91 by Jim Gillogly.
;
065V01 W9/VSCY9/89/-/8SCBAN/ZCP9/4OB9/VOE/SB89-
065V02 8AM/ZCCF/E/O89/-/9PCO/4OB*/AT/SCOPCC9/8AEAJ-  ; * may be R
065V03 PAE/SCOP/ZF9/9/-/*AR/SCO89-                   ; * may be S or R or 2??
065V04 POCC89/OPOR/ZC89/-/9PCC9/ZCO89/4OFZ8/9VSC9-
065V05 8ZC89/OFCO89/4OF8/-/ZX9/SOF9/SOFCO89/OFC9/89-
065V06 [49]O/9PSC9/ZXCW9/-/9FSC89/SC89/ZX89#         ; Probably 4 (?)

And now, with those conventions:

; Folio 65v (Herbal) appears as Plate II in Newbold.
; It features a plant with many roots, branches and flowers; the main
; stem separates the lines of text, which are well-registered across
; the stem.
; Transcribed 7 Dec 91 by Jim Gillogly.
;
{065V01} W9/ VSCY9/ 89/-/ 8SCBAN/ ZCP9/ 4OB9/ VOE/ SB89-
{065V02} 8AM/ ZCCF/ E/ O89/-/ 9PCO/ 4OB[*|R]/ AT/ SCOPCC9/ 8AEAJ-  
{065V03} PAE/ SCOP/ ZF9/ 9/-/ [*|S|R|2]AR/  SCO89-                   
{065V04} POCC89/ OPOR/ ZC89/-/ 9PCC9/ ZCO89/ 4OFZ8/ 9VSC9-
{065V05} 8ZC89/ OFCO89/ 4OF8/-/ ZX9/ SOF9/ SOFCO89/ OFC9/ 89-
{065V06} [4|9]O/ 9PSC9/ ZXCW9/-/ 9FSC89/ SC89/ ZX89#         

(Note that I have added spaces for readability) 
With that scheme, nothing prevents you from taking up several lines of transcription for just one line of Voynich, you may even put in your own translations! E.g.:


{065V05}    8  Z  C    8   9/ 
         {one who dies one s --> the ones who die like this one }
              O     F   C    O 8  9  / 
         { [h]o[c]  f[e]m[in]u n tur = are thereby feminated }

         { just having a go at Levitov and Brumbaugh, folks :-) }

More on alternative readings. 

It would not hurt to be able to express the probability of alternative readings (always optionally, of course), especially if we ever manage to put together some Voynich OCR algorithm. So here is an idea:

[4(95)|9(5)] would mean:  4: 95% probable, 9: 5% probable.

Now for the hard part: GIF. 

I mentioned GIF because it has fairly good compression ratios. JPEG is better, but not so appropriate for our purpose: JPEG is good for colour photographs, far less so for cartoons and writing. I'd say it would be better to have one GIF file per folio. Perhaps a data compression program such as the later version of LHA or ARJ could improve on GIF compression, given the type of data scanned (seat-of-the-pants feeling again). I have scanned and compressed most of the Easter Island tablets. Face A of Tablet Mamari, compressed with ARJ, comes to 75,560 bytes, face B to 75,976 bytes. Makes me think that a folio of Voynich would take anything from 30 to 100 Kbytes, compressed. 

And the hardest part is still to come.

I do not think any of the commercial (or PD, or shareware) graphic editors would serve our purposes too well. Just my experience in scanning and cleaning up the scanned images of the Easter Island tablets. I can help there only in writing something for PCs. Macs and Sun workstations I know nothing about worth mentioning. And I could use that software on the Easter Island tablets too (provided I ever get around to writing it!). We need to know who uses what machines. In the meantime it can't hurt of course to ferret about for some serviceable graphic editor. 

From gauss!rand.org!jim%mycroft Mon Dec 09 08:12:51 PST 1991
Received: by gauss; Mon Dec  9 12:01:00 EST 1991
Received: by inet.att.com; Mon Dec  9 12:00 EST 1991
Received: from mycroft.rand.org by rand.org; Mon, 9 Dec 91 08:12:55 -0800
Received: from localhost by mycroft.rand.org; Mon, 9 Dec 91 08:12:52 PST
Message-Id: <9112091612.AA05083@mycroft.rand.org>
To: voynich@rand.org
Subject: f 55v - independent readings
In-Reply-To: Your message of Mon, 09 Dec 91 09:42:21 -0500.
             <9112091442.AA05831@rand.org> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Mon, 09 Dec 91 08:12:51 PST
Sender: jim%mycroft@rand.org
Status: OR

Well, that was depressing.  Clearly it's going to be a long haul.
Does anybody else have Newbold?

I'll intersperse the Reeds and Gillogly transcriptions of this page, with
R on his and G on mine.


G <065V01> W9. VSCY9. 89.-. 8SCBAN. ZCP9. 4OB9. VOE. SB89-
R <65v.1>(W9.VSCW9.89)    {stem} (8SBAN.ZCP9.4OB9.VOE.SB89) {ghost 4 under 4 of
R 4OB9}

I'm right on the Y vs W in word 2 and the extra C in word 4.

G <065V02> 8AM. ZCCF. E. O89.-. 9PCO. 4OB[*|R]. AT. SCOPCC9. 8AEAJ-
R <65v.2>(8AM.ZCCF,E,O89)   {stem} (9PCO.4OB2.AT.SCOPCCO.8AEOJ)

I wasn't doing commas for possible or probable spaces, so that doesn't
count as a difference yet.  His reading of the 2 wasn't one of my choices,
but is credible -- it seems to have an angle on the bottom of the sickle,
though.

The 9 vs O and A vs O in the last two words are interesting.  I suspect
we could resolve this with a better print.

G <065V03> PAE. SCOP. ZF9. 9.-. [*|S|R|2]AR.  SCO89-
R <65v.3>(9PAE.SCOPZF9.9)   {stem} (2A2.SCO89) {initial 2 of 2A2 doubtful}

He's right on the 9 in the first word.  I'd put at least a comma between
SCOP and ZF9.  After the stem, the old 2/R problem again.  I guess I believe
his 2nd 2, but am still undertain about the first atypical one.

G <065V04> POCC89. OPOR. ZC89.-. 9PCC9. ZCO89. 4OFZ8. 9VSC9-
R <65v.4>(POCC89.OPOR,ZC89) {stem} (9PCC9.ZCO89.4OFZ8.9VSC9)

My god, we agreed on one!

G <065V05> 8ZC89. OFCO89. 4OF8.-. ZX9. SOF9. SOFCO89. OFC9. 89-
R <65v.5>(8ZC89.OFCO89.4OF8){stem} (ZX9.SOF9.SOFCO89.OFC989)

And another (barring the last word break).

G <065V06> [4|9]O. 9PSC9. ZXCW9.-. 9FSC89. SC89. ZX89#
R <65v.6>(4O.9PSC9.ZXCW9)   {stem} (9FSC89.SC89.ZCFC89)

In the last word here I'm seeing the C's connected through the F, making
the CFC into an X.

Clearly the line comparer that I alluded to earlier won't nearly be up to
this kind of comparison.  Time for some tool-building?

This is a lot of variation for a reasonably-clear print of a six-line
page!

Jim Gillogly

From gauss!math.mit.edu!jbaez Mon Dec  9 11:34:59 EST 1991
Received: by gauss; Mon Dec  9 12:08:22 EST 1991
Received: by inet.att.com; Mon Dec  9 12:08 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Mon, 9 Dec 91 08:37:42 -0800
Received: from cayley (CAYLEY.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA14410; Mon, 9 Dec 91 11:33:46 EST
Received: by cayley; Mon, 9 Dec 91 11:34:59 EST
Date: Mon, 9 Dec 91 11:34:59 EST
From: jbaez@math.mit.edu
Message-Id: <9112091634.AA19744@cayley>
To: voynich@rand.org
Status: OR

Okay, I'll continue to pester Yale avidly concerning a photocopy.
It's odd how few people have actually seen even of a photocopy of the
whole thing!  D'Imperio says sheworked from 4th-generation photocopies
that were covered with scrawls of previous people's work.

I have the address of who to contact at Yale, in case anyone wants it,
but for now you may all assume I'm taking care of that.  Please don't
deluge the poor guy with requests for photocopies, as I may be able to 
provide them more cheaply once I get them.

The address is


	Robert Babcock
	1603a Yale Station
	New Haven, CT  06520

	
I had written:

 Once we have such standardized access to the
 real thing, high-tech solutions involving GIF's etc. would seem
 unnecessary.

To which Jim responded:

 I disagree.  If we can afford the size of GIFs they will make it a lot
 easier
 for us to communicate about specific parts -- clip a woman in a vat with
 the
 caption, rather than trying to describe where in our jointly-shared
 hard-copy archive it is.  Also, note that there are now 15 people on the
 list.  Distributing hard-copy stuff is going to get prohibitive very fast!

What I really meant was that transcription should not require access to 
GIF capability - I certainly have no objection to using GIFs as a means
of communication (except for the fact that I've never been able to get
GIFs to work, which is purely my own fooolishness).


From gauss!BINAH.CC.BRANDEIS.EDU!EVANS Mon Dec  9 14:57 EDT 1991
Received: by gauss; Mon Dec  9 15:30:52 EST 1991
Received: by inet.att.com; Mon Dec  9 15:30 EST 1991
Received: from binah.cc.brandeis.edu by rand.org; Mon, 9 Dec 91 11:59:06 -0800
Received: from BINAH.CC.BRANDEIS.EDU by BINAH.CC.BRANDEIS.EDU (PMDF #12080) id
 <01GDWW0EDQ9C9S43VZ@BINAH.CC.BRANDEIS.EDU>; Mon, 9 Dec 1991 14:57 EDT
Date: Mon, 9 Dec 1991 14:57 EDT
From: Ronald Hale-Evans <EVANS@BINAH.CC.BRANDEIS.EDU>
Subject: NSA
To: voynich@rand.org
Message-Id: <01GDWW0EDQ9C9S43VZ@BINAH.CC.BRANDEIS.EDU>
X-Vms-To: VOYNICH
X-Vms-Cc: EVANS
Status: OR

Greetings. I remember reading somewhere that the National Security Agency had
analysed the Ms. (tho. not necessarily deciphered it). I don't remember the
source of this information, and of course, like all information about the NSA,
it could be disinformation. Does anyone have any more (dis)information on this
topic?

I have a copy of Brumbaugh's *The Most Mysterious Manuscript* and frankly, I
think it's hogslop. His methods remind me of people trying to find alphabetic
mnemonics for telephone numbers; using the same methods, I can prove that it's
really a vegetarian cookbook written in the late '60s. I went to Yale for my
undergraduate studies and have visited the Beinecke more than once. I have
never laid hands on the Ms. proper (don't ask me why; I had the privileges),
but they gave me some photocopies that I can duplicate for people. If anyone
has any trouble accessing the Ms. at the Beinecke, feel free to contact me. I
believe that Yale may grant Beinecke access to alums for a fee.

Reading through the digests on rand.org, I notice that someone mentioned the
Codex Seraphinianus and asked if anyone had done any work on it. I have managed
to largely decode the numbering system at the bottom of the pages. It is not a
number-place system like Arabic or binary. It works more like Roman numerals. I
can supply some more information if anyone wishes.

I have read other books of William Poundstone's, but not *Labyrinths of
Reason*. It does not surprise me that he devotes a chapter of one of his books
to the Ms.. Could someone email me privately with a bit more info about the
other topics covered by the book? I am also interested in finding a
glossary/grammar of Enochian. Can someone supply me with some pointers?

As for an introduction, I work as a user consultant at Brandeis University. I
have been interested in cryptology since I was a child, but some of the more
abstruse stuff like the RSA algorithm has slipped by me. To be honest, I have
no real experience in computer analysis of text, either ciphertext or
plaintext, but I'm willing to work with and learn from people in the Boston
area who do. I am a contributor the the Constructed Languages mailing list,
where I have posted information about this list. (If you would like to join that
list, send email to conlang-request@buphy.bu.edu. I believe at least one of the
Voynich postings from sci.lang was posted there; there should be some overlap
of interests.) 

Ron Hale-Evans                                                               

From gauss!trl.OZ.AU!j.guy Mon Dec  9 15:17:24 EST 1991
Received: by gauss; Sun Dec  8 23:50:59 EST 1991
Received: by inet.att.com; Sun Dec  8 23:50 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sun, 8 Dec 91 20:18:41 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA08963; Mon, 9 Dec 91 15:17:24 EST
Date: Mon, 9 Dec 91 15:17:24 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112090417.AA08963@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: equipment, ground rules
Status: OR


My equipment:

Sun Sparcstation IPC, connected to postscript laser printer.
PC Clone 386SX 16MHz, Phoenix BIOS, 640K RAM, 85 Meg HD, EGA.
PC Clone 386DX 33MHz, AMI BIOS, 8 Meg RAM, 85 Meg HD, SVGA with 512K Paradise card.

Sources:

D'Imperio
Herbal with 2 color reproductions from Voynich
about 2 dozen black-and-white 10x18 enlargements made from Yale microfilm
the 2 folios in Bennett's book (That's all folks!)

I am very fluent in Turbo Pascal and OOP.


A Voynich font for display on PC should be dead easy. But for nice printing 
we'd need a stroked font, though. Harder.

Back to ground rules:

Jim Reeds' suggestion of using "." for word spaces. Why, of course, that's ideal! Unobtrusive, perfect. Comma for doubtful break? Yes, it's a good mnemonic. Excellent.

The points made about page numbers being comments but comments that matter is right. We need different brackets for those. Wedges, as suggested by Jim Gillogly, are fine. They're also easy to reach on the keyboard and that's important. Above all, the system must be human-readable. For that reason, I would be against allowing for regular expressions "a la grep". They may be grep-friendly, but get downright human-hostile in no expression flat. Must not only keep it, but work hard at making it simple. It is true that Gilles (a male in French, not like Jill) might transcribe (bcde|af|ade|bcf) what this Jim (Jim==James==Jacques via Giacomo) transcribes (a|bc)(de|f), but all we need is lay down a few very simple transcription guidelines to prevent that happening. (You'll have to give me another week-end to come up with them in simple English, though. Saying it simply is not simple).

Jim Gillogly wonders how to handle "hand" and "language". So do I.... We're soon going to run out of brackets if we are extravagant with them. 

After some thinking, I suggest that handwriting should be signalled as it is now, by a single letter (A or B) appended to the folio and line identifier. When unknown, or doubtful, nothing appended. Should the handwriting change within a line, insert a comment <A> or <B>. As for comments about the language, they really are comments about the alphabet used. By default, we consider the alphabet to be that of the "Voynich language". Where it switches to say, what we think is 16th century German we insert a comment, <German> or something like that, and where it goes back to Voynich <Voynich> or just <V>, or <A> or <B> if we could identify which handwriting it is. The rule would be:

   If it starts with a digit, it's a folio and line-number comment.
   If it is a single letter, it is a Voynich/handwriting comment.
   If it starts with a letter, and there's more stuff following, it is a "language" comment.

For instance:

{ Hitherto unknown folio discovered in my Xmas stocking, 25/12/1991 5:23 am Melbourne
  local time. Arbitrarily numbered 666R }

<666R01A> W9. VSCY9. 89.-. 8SCBAN. ZCP9. <B> 4OB9. VOE. SB89-
<666R02B> 8AM. ZCCF. <A> E. <Enochian> Napeai babagen, ds
<666R03>  brin vx ooaona <V> PAE. SCOP. <Didot 6 pts> Ave 
<666R04>  Caesar, moritura in homage to Levitov te salutat! <A> 9FSC89.
<666R05A> ZX89#

Actually, it's better to say that those "language comments" are really "font comments", for that is what they are.


Jacques Guy 

From gauss!G.GP.CS.CMU.EDU!Karl.Kluge Mon Dec  9 23:50 EST 1991
Received: by gauss; Tue Dec 10 00:34:24 EST 1991
Received: by inet.att.com; Tue Dec 10 00:34 EST 1991
Received: from G.GP.CS.CMU.EDU by rand.org; Mon, 9 Dec 91 21:27:46 -0800
Date: Mon, 9 Dec 1991 23:50-EST
From: Karl.Kluge@G.GP.CS.CMU.EDU
To: voynich@rand.org
Subject: Re: NSA
Message-Id: <692340600/kck@G.GP.CS.CMU.EDU>
Status: OR

(By the way, as a quick intro, I am a doctoral student in computer
science at Carnegie Mellon. My research area is outdoor mobile robot
navigation, specifically my system uses color road images to drive
the NAVLAB test vehicle. I'm also getting into thesis crunch, which
may limit my ability to spend much time helping with transcription.)

> Date: Mon, 9 Dec 1991 14:57 EDT
> From: Ronald Hale-Evans <EVANS@BINAH.CC.BRANDEIS.EDU>
> Subject: NSA
> 
> Greetings. I remember reading somewhere that the National Security Agency 
> had analysed the Ms. (tho. not necessarily deciphered it). 

The NSA put out at TR on the Mss. "The Voynich Manuscript: An Elegant
Enigma", by M. E. D'Imperio.

> I have a copy of Brumbaugh's *The Most Mysterious Manuscript* and frankly,
> I think it's hogslop. His methods remind me of people trying to find 
> alphabetic mnemonics for telephone numbers; using the same methods, I can
> prove that it's really a vegetarian cookbook written in the late '60s.

Brumbaugh may well be wrong, but I think "hogslop" is overly harsh. In
my notes on occult alphabets I have one attributed to Cornelius Agrippa
which uses 9 characters, mapping multiple letters to each ("]", for 
instance, is [final-m, c, v, or oo]; "[" is [th, m, or d]; etc.) I copied the
page from D'Imperio with the supposed cryptarithmatic, but haven't had
a chance to look at it yet. Most office phone systems with directories
use essentially the system Brumbaugh suggests, having you key the name
in on the phone keypad -- I suspect the ambiguity problem is not as bad
as you think.

> I am also interested in finding a glossary/grammar of Enochian. 

I don't know of published vocab/grammar, but you should be able to work
it out for yourself given the text of the calls. Regardie's GOLDEN DAWN
has 19 of them, as does ELIZABETHAN MAGIC by Robert Turner. Here are the
first three (without their English translation):

Ol sonf vorsg, goho Iad balt lansh calz vonpho, sobra z-ol ror i ta Nazpsad
Graa to Malprag Ds hol-q qaa nothoa zimz Od commah ta nobloh zien: Soba
thil gnonp prge aldi Od vrbs oboleh grsam Casarm ohorela caba pir Od
zonsrensg cab erm Iadnah Pilah farzm zurza adna Ds gono Iadpil Ds hom
Od toh Soba Ipam lu Ipamis Ds loholo vep zomd Poamal Od bogpa aai ta
piap piamo-i od vaoan ZACARe c-a od ZAMRAN Odo cicle Qaa Zorge, Lap
zirdo Noco MAD Hoath Iaida.

Adgt v-pa-ah zongom fa-a-ip Sald vi-i-v L sobam I-al-prg I-za-zaz
pi-adph cas-arma abramg ta talho paracleda Q-ta lors-l-q turbs ooge
Baltch Giui chis lusd orri Od mi-calp chis bia ozongon Lap noan trof
cors tage o-q manin Ia-i-don Torzu gohel ZACAR ca c-no-qod, ZAMRAN
micalzo Od ozazm vrelp Lap zir Ioiad.

Micma goho Piad zir com-selh a zien blab Os Lon-doh Norz Chis othil
Gigipah vnd-l chis ta-pu-im Q mos-pleh teloch Qui-i-n tolorg chis i
chis ge m ozien dst brgda od torzul i li F ol balzarg, od aala Thinl
Os ne ta ab dluga vomsarg lonsa cap-mi-ali vors cla homil cocasb
fafen izizop od mi i noag de gnetaab vaun na-na-e-el panpir Malpirgi
caosg Pild noan vnalah balt od vooan do o-i-ap MAD Goholor gohus
amiran Micma Iehusoz ca-ca-com od do-o-a-in noar mi-ca-olz a-ai-om
Casarmg gohia ZACAR vniglag od Im-ua-mar pugo plapi ananael Q a an.

Note the differences between Turner's transcription of the third call
and the transcription in fig 44 of D'Imperio.

Karl

From gauss!rand.org!jim%mycroft Mon Dec 09 18:57:53 PST 1991
Received: by gauss; Mon Dec  9 22:30:57 EST 1991
Received: by inet.att.com; Mon Dec  9 22:30 EST 1991
Received: from mycroft.rand.org by rand.org; Mon, 9 Dec 91 18:57:57 -0800
Received: from localhost by mycroft.rand.org; Mon, 9 Dec 91 18:57:54 PST
Message-Id: <9112100257.AA01828@mycroft.rand.org>
To: voynich@rand.org
Subject: GIF image of Voynich f 3v
Date: Mon, 09 Dec 91 18:57:53 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

I've put a GIF of the text part of f 3v in the ftp area, pub/jim/f3v.gif
on rand.org, for those who would like to try out their image manipulation
tools on it.  This is about 35K, to see whether this is going to be an
adequate sampling rate for real work.  It was printed from my microfilm on
a somewhat wimpy printing reader, scanned at 300 dpi, reduced to about 150
dpi, and the gray scale was reduced as well.  It looks a *whole* *lot*
better in the 1 MB file, which is too big for our ftp directory.

I suspect other pages with more text will need larger GIF files to be
legible.

Jim Gillogly

From gauss!ns1.CC.Lehigh.EDU!MAILER-DAEMON Tue Dec 10 09:34:32 0500 1991
Received: by gauss; Thu Dec 12 14:51:35 EST 1991
Received: by inet.att.com; Thu Dec 12 14:51 EST 1991
Received: from rand.org by ns1.CC.Lehigh.EDU (AIX 3.1/UCB 5.61/4.03)
          id AA19213; Tue, 10 Dec 91 09:34:32 -0500
Date: Tue, 10 Dec 91 09:34:32 -0500
From: MAILER-DAEMON@ns1.CC.Lehigh.EDU (Mail Delivery Subsystem)
Subject: Returned mail: unknown mailer error 3
Message-Id: <9112121950.AA19213@ns1.CC.Lehigh.EDU>
To: <reeds@gauss.att.com>
Status: OR

   ----- Transcript of session follows -----
bellmail: error closing file: No space left on device
554 <mlf3@ns1.cc.lehigh.edu>... unknown mailer error 3

   ----- Unsent message follows -----
Received: from rand.org by ns1.CC.Lehigh.EDU (AIX 3.1/UCB 5.61/4.03)
          id AA59660; Thu, 12 Dec 91 14:50:33 -0500
Received: from research.att.com by rand.org; Thu, 12 Dec 91 10:39:38 -0800
Message-Id: <9112121839.AA29330@rand.org>
From: reeds@gauss.att.com
Date: Thu, 12 Dec 91 13:39:21 EST
To: voynich@rand.org

Salvete!

A much corrected list of folios and images in rand.org:pub/jim/foliations,
thanks to data supplied by Karl Kluge.  (And thanks to technology change:
I went over to using 3 by 5 cards, which are much more efficient for this
task than anything on the computer!)

Here is a breakdown of the 232 ``pages'' I think exist:
					
					In D'Imp's transc.	Not in transc.
Image Published					16			18
Image not published but on Jim G's BM film	103			0
Image not published nor on BM film.		12			83

Totals						131			101

We have transcribed some (maybe half a dozen) pages so far, but I regard
them as unchecked drafts or practice runs.  Of course the Beinecke
film covers the whole MS, but we don't have a copy of that yet.  And
some of the forms of ``publication'' listed above are pretty obscure,
as are some of the images.

Assuming that we transcribe each of D'Imps pages once, and each of the
others twice, the bare minimum needed to get a check on everything, we
have 333 pages to go: halfway to perdition!

Merry Michiton, and a happy Oladabas Day to you!

Jim Reeds

From gauss!rand.org!jim%mycroft Tue Dec 10 06:37:42 PST 1991
Received: by gauss; Tue Dec 10 09:55:21 EST 1991
Received: by inet.att.com; Tue Dec 10 09:55 EST 1991
Received: from mycroft.rand.org by rand.org; Tue, 10 Dec 91 06:37:45 -0800
Received: from localhost by mycroft.rand.org; Tue, 10 Dec 91 06:37:43 PST
Message-Id: <9112101437.AA02551@mycroft.rand.org>
To: voynich@rand.org
Subject: Periods instead of slashes in current working text
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Tue, 10 Dec 91 06:37:42 PST
Sender: jim%mycroft@rand.org
Status: OR

Looks like general appropbation for using <> on line numbers and '.'
and ',' instead of '/' for white space.  I've modified the current
working version, called "voynich.now" in the ftp directory to take
these into account.  The text now looks like this:

<00101A> VAS92.9FAE.AR.APAM.ZOE.ZOR9.QOR92.9.FOR.ZOE89-
<00102A> 2OR9.XAR.O.R.9.FAN.ZPAM.ZAR.AR*.QAR.QAR.8AD-
<00103A> 29AU.ZCF9.OR.9FAM.ZO8.QOAR9.Q*R.8ARAM.29-
<00104A> $OM.OPCC9.OPCOR.2OEOP9.Q*AR.8AM.OFAM.OE.OFAD-
<00105A> 2AT.9.SCAR.QAM.WAR.YAM#
<00106A> 98ARAIZO#

Jim Reeds implicitly proposed parens for line segments, and gave
examples, e.g.:

> <65v.2>(8AM.ZCCF,E,O89)   {stem} (9PCO.4OB2.AT.SCOPCCO.8AEOJ)
> <65v.3>(9PAE.SCOPZF9.9)   {stem} (2A2.SCO89) {initial 2 of 2A2 doubtful}

I'm not convinced this is valuable.  Currier has shown that the line
has an important function (statistics are different at beginning and end
of lines), but I don't know whether that includes the segments on a single
line, like one separated by a plant stem, and I don't think our notation
should anticipate the answer.

On the other hand, I'm also not wild about using the end-of-line character
in the middle of a line, as D'Imperio does.  That's making the same
judgement, and if we go to a multiple line format for special investigations
(cf Jacques Guy's translation example), it becomes unclear where the
physical line has ended.

And yes, I have a proposed solution: double dot ".." for a wide gap
within a line.  I'd propose to change all internal dashes in the D'Imperio
text to double dots, and leave the final dash as-is.

Comments?

Jim Gillogly

From gauss!BINAH.CC.BRANDEIS.EDU!EVANS Tue Dec 10 11:18 EDT 1991
Received: by gauss; Tue Dec 10 12:14:18 EST 1991
Received: by inet.att.com; Tue Dec 10 12:14 EST 1991
Received: from binah.cc.brandeis.edu by rand.org; Tue, 10 Dec 91 08:20:02 -0800
Received: from BINAH.CC.BRANDEIS.EDU by BINAH.CC.BRANDEIS.EDU (PMDF #12080) id
 <01GDY2NI7R6O9S49N4@BINAH.CC.BRANDEIS.EDU>; Tue, 10 Dec 1991 11:18 EDT
Date: Tue, 10 Dec 1991 11:18 EDT
From: Ronald Hale-Evans <EVANS@BINAH.CC.BRANDEIS.EDU>
Subject: my photocopied pages
To: voynich@rand.org
Message-Id: <01GDY2NI7R6O9S49N4@BINAH.CC.BRANDEIS.EDU>
X-Vms-To: VOYNICH
X-Vms-Cc: EVANS
Status: OR

For those who want copies of my photocopies: I'll be happy to snailmail them to
you free of charge. Just make sure I have your address. In return, I ask that
you be patient. The holiday craziness is coming up and since I'mm remodeling my
library, a lot of my papers are inaccessibly packed away.

A couple of details: I only have a few random pages of photocopies. I haven't
checked tham against Brumbaugh or anything, so for all I know they could be
duplicated in available materials. As for other appearances of Ms. pp., I
remember seeing at least one large, full-colour page in a popular book on
cryptography I read when a teenager. Since I'm going to my parents' house for a
visit soon, I'll see if I can track this down at the local library.

Ron

From gauss!gauss.att.com!reeds Tue Dec 10 12:09:03 EST 1991
Received: by gauss; Tue Dec 10 12:43:12 EST 1991
Received: by inet.att.com; Tue Dec 10 12:43 EST 1991
Received: from research.att.com by rand.org; Tue, 10 Dec 91 09:09:20 -0800
Message-Id: <9112101709.AA06147@rand.org>
From: reeds@gauss.att.com
Date: Tue, 10 Dec 91 12:09:03 EST
To: voynich@rand.org
Status: OR

About parens for text; speculations on lines in general.

Jim Gillogly writes

> Jim Reeds implicitly proposed parens for line segments, and gave
> examples, e.g.:
> 
> > <65v.2>(8AM.ZCCF,E,O89)   {stem} (9PCO.4OB2.AT.SCOPCCO.8AEOJ)
> > <65v.3>(9PAE.SCOPZF9.9)   {stem} (2A2.SCO89) {initial 2 of 2A2 doubtful}
> 
> I'm not convinced this is valuable.  Currier has shown that the line
> has an important function (statistics are different at beginning and end
> of lines), but I don't know whether that includes the segments on a single
> line, like one separated by a plant stem, and I don't think our notation
> should anticipate the answer.

Well, I'm not either, 100%; I was experimenting.

If we do NOT record location of wide gaps in lines, we will never be
able to gather statistics to settle the matter.  If we DO record location
of wide gaps, we can always ignore that information.  The precise notation 
for a wide gap is not important.  But when there is an obvious feature
cutting across several lines, creating real doubt about whether its one
``column'' of text or two, I think we should note it, with a notation
that's distinguishable from the notation we might use for a slightly wider
than average word space such as may occur here & there without pattern.
I have no objection to .. as long as its reserved for possible column-
separating gaps, and not just for     the occasional isolated wide inter-
word gap like the one just above HERE.  Can you tell from your BM film
(you got it from BM, right, but now they call themselves BL, so what
do WE call it?), Jim, how D'Imperio uses internal - ?

I suspect that final - as a line ender is some kind of IBM card artifact 
connected with treatment of 'continuation cards'.

Now for the speculation.

This whole business of line statitics is intriguing.  I found that the 
distribution of initials of second words in lines is also different 
from the distribution of initials of generic words.  If the line
represents the ``compositional unit'' of the text you would expect
line-based regularities, but not otherwise.  If the MS were composed
line by line as it was being written you would expect line-based 
regularities.  Modern examples of writing exhibiting coherence between
compositional unit and writing line are poetry, dialogue in fiction,
and in indexes, but not in general purpose prose.  (But now I remember
the T. C. Mits books.  God, they were weird!)  I don't think things 
were different in this regard in the 16th century.

Here is a thought question: Are the stem-broken lines in 33v/34r
continuous lines of text?  (``Sunflowers'', plate in Kahn, near p844,
also Blunt & Rafael,  p89/90.) The alignment is not exact, and the line
parts were clearly not written at the same time (you can see when the 
pen was dipped).  This argues for either:

(1) copying from a draft whose text was already laid out into lines, with 
the copyist NOT following reading-order. (Maybe a professional scribe 
copying text he cannot understand?  Jacques's comments aside, the writing
looks extraordinarily neat and tidy to me.  My parents have a charter 
from Rudolph II's court hanging in their dining room; it is not nearly 
as legibile as VMS.  Of course its in chancery hand, written by & for
lawyers, so maybe it doesn't signify...)

or

(2) the text just there is cover text, ``fill'', if you will, produced
by whatever means the rest of the text was produced by.  This would not
preclude acrostics, but would, I think, rule out Baconian-Baudot 
``biliteral'' or Dalgarnoid artificial languages, etc, AT THAT LOCUS.

I have noticed strong regularities between word ending letter and initial
of next word; it shoud be easy to see if they hold up for stem-straddling
word pairs.  This will be my homework.

Changing the subject.  Does anyone know what ff are illustrated in
Kraus's 35 MSS?  If you do, or can easilly find out, let me know.

Jim Reeds

From gauss!anubis.network.com!dean Tue Dec 10 12:02:26 CST 1991
Received: by gauss; Tue Dec 10 13:23:53 EST 1991
Received: by inet.att.com; Tue Dec 10 13:23 EST 1991
Received: from nsco.network.com by rand.org; Tue, 10 Dec 91 10:06:47 -0800
Received: from anubis.network.com by nsco.network.com (5.61/1.34)
	id AA10703; Tue, 10 Dec 91 12:07:25 -0600
Received: from ns.network.com by anubis.network.com (4.0/SMI-4.0)
	id AA12362; Tue, 10 Dec 91 12:02:26 CST
Date: Tue, 10 Dec 91 12:02:26 CST
From: dean@anubis.network.com (Dean C. Gahlon)
Message-Id: <9112101802.AA12362@anubis.network.com>
Received: by ns.network.com (4.1/SMI-4.1)
	id AA27124; Tue, 10 Dec 91 12:02:25 CST
To: EVANS@BINAH.CC.BRANDEIS.EDU
Cc: voynich@rand.org
In-Reply-To: Ronald Hale-Evans's message of Tue, 10 Dec 1991 11:18 EDT <01GDY2NI7R6O9S49N4@BINAH.CC.BRANDEIS.EDU>
Subject: my photocopied pages
Status: OR

If you'd care to send my copies of your photocopies of the pages from
the Voynich MS., I'd appreciate it very much. My snailmail address is:

	Dean C. Gahlon
	3553 Pleasant Ave. S.
	Minneapolis, MN 55408

Thanks.

From gauss!math.mit.edu!jbaez Tue Dec 10 13:45:53 EST 1991
Received: by gauss; Tue Dec 10 14:00:01 EST 1991
Received: by inet.att.com; Tue Dec 10 13:59 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Tue, 10 Dec 91 10:47:53 -0800
Received: from riesz (RIESZ.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA19599; Tue, 10 Dec 91 13:43:11 EST
Received: by riesz; Tue, 10 Dec 91 13:45:53 EST
Date: Tue, 10 Dec 91 13:45:53 EST
From: jbaez@math.mit.edu
Message-Id: <9112101845.AA12035@riesz>
To: voynich@rand.org
Status: OR

The following speculation of Jim's is fascinating:

This whole business of line statistics is intriguing.  I found that the 
distribution of initials of second words in lines is also different 
from the distribution of initials of generic words.  If the line
represents the ``compositional unit'' of the text you would expect
line-based regularities, but not otherwise.
------  

Could you give us more precise info on how strong the statistical
difference of initial and second words in lines are from generic words?
Nate and I are going to get some statistical analyses running but for
now I'll just throw out some questions and hypotheses.  

Are there strong statistical differences in end-of-line words?  This
would go along with rhyming, or partially rhymed poetry.  I have 
trounble understanding what would give beginning-of-line peculiarities
except:

1)  Writing backwards.  Of course the neat left margins argue against
this but any smart and sufficiently hardworking author... why don't we
call him, her, or them "X"... could realize this and deliberately
produce neat left margins.  If there was rhyming poetry one should find
n-line periodicities as well (n = 2 for couplets and so on).  We should
check this.

2)  An "index" or "dictionary" of some sort that goes "Term, information
about term".  Here one would expect there to be less repetition of 
initial words than of generic words (think of a dictionary, for
example).  Of course there could be more than one line in the
"definition".  (One should think not just of dictionary, of course, but
of more general texts with this structure: herbals, encyclopedias, 
mail-order catalogs :-), etc.)  One might find *less* entropy in the
second word becuase definitions (or whatever) tend to begin "This ..."
or "A...".

3)  A kind of incantation, epic or poem -- let's just say a text with
more rhythmic structure than ordinary prose -- that relies on
beginning-of-line effects rather than usual rhyme.  Here, as with 1),
one would expect *less*, rather than more entropy among
beginning-of-line words than generic words.

4)  The Voynich was nonsense written by X in a line-by-line manner,
which naturally gives the first couple of words in a line different
statistical properties.

5)  Most of the text is cover, or fill, and these statistical properties
are not the key to the real *meaning* of the text.


So I guess my key quantitative question is whether the entropy of
initial words (and here I mean entropy on the word-by-word level, rather
than entropy on the letter-by-letter or digraph-by-digraph level) is
higher or lower than for generic words; same question for second words
and last words.

This observation by Jim --- if it's statistically significant! --- is
exciting and pleasing becuase it shows that perhaps by careful
statistical analysis of the text we can can determine a lot of useful
things about its structure.  Here it's important to keep two aspects
straight: on the one hand, the formation and testing of hypotheses,
which will no doubt be a roller-coaster of victories and defeats, and on
the other hand the collection of statistical data about the text, which
will gradually accumulate and stand as real progress regardless of the
flux of hypotheses.



Jim also raises another interesting point:

I have noticed strong regularities between word ending letter and initial
of next word; it shoud be easy to see if they hold up for stem-straddling
word pairs.  This will be my homework.
-------
Again this is odd, and seems to argue against any normal sort of text.
Again, quantitative measurement of how strong this effect is would be
interesting.  

Good going, Jim!

John Baez



From gauss!gauss.att.com!reeds Tue Dec 10 14:52:27 EST 1991
Received: by gauss; Tue Dec 10 15:22:30 EST 1991
Received: by inet.att.com; Tue Dec 10 15:22 EST 1991
Received: from research.att.com by rand.org; Tue, 10 Dec 91 11:54:53 -0800
Message-Id: <9112101954.AA09570@rand.org>
From: reeds@gauss.att.com
Date: Tue, 10 Dec 91 14:52:27 EST
To: voynich@rand.org
Status: OR

1. I looked a bit at the GIF file Jim Gillogly scanned.  Looks almost
clear enough to be a PLEASURE to transcribe from.  I understand there are
such things as microfilm scanners: in goes a 35mm reel, out comes
a gigabit of crisp image data.  I'm converted!  As for developing graphics
tools to aid in the transliteration, I'm not so sure.

2. I will do my statistics homework and supply John Baez and the list with
word ending statistics, & the like.  (I've got lots of notes scattered here
and there, lots of little tabulate this/caluculate that sorts of programs,
and have got to find the right one to look at again.)

3. A new version of my list is ready, entitled

	A Provisional Checklist of Voynich MS ``pages''

which I will send to Jim Gillogly for placing in the ftp loading dock.
If you see errors or can add in any way to to, please let me know by
direct email to reeds@research.att.com.

Jim Reeds

From gauss!rand.org!jim%mycroft Tue Dec 10 12:02:45 PST 1991
Received: by gauss; Tue Dec 10 15:35:15 EST 1991
Received: by inet.att.com; Tue Dec 10 15:35 EST 1991
Received: from mycroft.rand.org by rand.org; Tue, 10 Dec 91 12:02:48 -0800
Received: from localhost by mycroft.rand.org; Tue, 10 Dec 91 12:02:45 PST
Message-Id: <9112102002.AA03368@mycroft.rand.org>
To: voynich@rand.org
Subject: The manuscript was copied
Date: Tue, 10 Dec 91 12:02:45 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

The morphological observations that lines can mis-register across a stem
and that rows are flush against the left margin may not be especially
relevant *in themselves* in view of the fact that it was all re-copied.

As Jim points out, the text need not have been copied in the same order
in which it was originally written, leading to blocks in the fork of
a double stem being written together, for example.  If written originally
right to left and copied left to right, the text can be flush left.

It seems clear that it was in fact copied, because of the lines in which
you have a couple of S characters widely separated with a very wide P
coming down into the middle of each of them.  This looks a whole lot like
a pair of Q's that the copyist was having fun with... you don't know
that you'll have the opportunity to play this game (and hence avoid finishing
the first Q) unless you see the text in front of you.

Jim Gillogly

From gauss!castrov.cuc.ab.ca!wuth Tue Dec 10 15:32:08 MDT 1991
Received: by gauss; Tue Dec 10 18:46:10 EST 1991
Received: by inet.att.com; Tue Dec 10 18:46 EST 1991
Received: from fsa.cpsc.ucalgary.ca by rand.org; Tue, 10 Dec 91 14:59:10 -0800
Received: by fsa.cpsc.ucalgary.ca (4.1/CSd1.2)
	id AA27921; Tue, 10 Dec 91 15:57:28 MST
Return-Path: <wuth@castrov.cuc.ab.ca>
Received: from castrov.cuc.ab.ca by ajfcal.cuc.ab.ca (5.65c/Cuc2.2)
	id AA11869; Tue, 10 Dec 1991 15:45:48 -0700
Received: by castrov.cuc.ab.ca (V1.13/Amiga)
	id AA03603; Tue, 10 Dec 91 15:32:08 MDT
Date: Tue, 10 Dec 91 15:32:08 MDT
Message-Id: <9112102132.AA03603@castrov.cuc.ab.ca>
From: wuth@castrov.cuc.ab.ca (Brett Wuth)
To: voynich@rand.org
Subject: Re: beginning-of-line
Status: OR

Please forgive me if I express my niavety.  I just joined this list.

<jbaez@math.mit.edu> writes:
>I have 
>trounble understanding what would give beginning-of-line peculiarities
>except:
>
>1)  Writing backwards.

>2)  An "index" or "dictionary" 

>3)  A kind of incantation, epic or poem

>4)  The Voynich was nonsense written by X in a line-by-line manner,

>5)  Most of the text is cover, or fill, and these statistical properties
>are not the key to the real *meaning* of the text.

I would add:

6) The text is written in a line-by-line manner in language with a
positional-coded grammar (or even style).  English is an example.
We tend to place the subject at the start of the sentence.  So the
initial word has a different distribution than the entire text.
I seem to remember the style in classical Latin was to place the
verb first.  Add a grammar that encodes function in the form of the
word (as almost all languages except English do) and you would get
even stronger correlations.

--
Brett Wuth   (403) 242-0848
wuth@castrov.cuc.ab.ca      BCWuth@uncamult.bitnet     wuth@castrov.uucp

From gauss!G.GP.CS.CMU.EDU!Karl.Kluge Tue Dec 10 21:47 EST 1991
Received: by gauss; Tue Dec 10 22:30:22 EST 1991
Received: by inet.att.com; Tue Dec 10 22:30 EST 1991
Received: from G.GP.CS.CMU.EDU by rand.org; Tue, 10 Dec 91 18:53:08 -0800
Date: Tue, 10 Dec 1991 21:47-EST
From: Karl.Kluge@G.GP.CS.CMU.EDU
To: voynich@rand.org
Subject: Equiptment & resources
Message-Id: <692419652/kck@G.GP.CS.CMU.EDU>
Status: OR

I have a Sun4 running UNIX on my desk which can print on various
departmental laser printers. Fluent in C, but no existing crypto code.

CMU or Pitt libraries have D'Imperio, Brumbaugh, THE ILLUSTRATED
HERBAL, the dealer catalog from when it was sold (with a B&W
reproduction of the fold-out folios as well as B&W photos of several
other folios).

Karl

From gauss!rand.org!jim%mycroft Tue Dec 10 13:57:19 PST 1991
Received: by gauss; Tue Dec 10 17:21:06 EST 1991
Received: by inet.att.com; Tue Dec 10 17:21 EST 1991
Received: from mycroft.rand.org by rand.org; Tue, 10 Dec 91 13:57:23 -0800
Received: from localhost by mycroft.rand.org; Tue, 10 Dec 91 13:57:20 PST
Message-Id: <9112102157.AA03774@mycroft.rand.org>
To: voynich@rand.org
Subject: Voynich transcriber volunteers?
Date: Tue, 10 Dec 91 13:57:19 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

While we're waiting for the Beinecke prints, let's start transcribing.  We
have a number of pages available for entering (mostly Newbold) or checking
(mostly BL).  I think we're converging on a format, so by the time you get
your pages in the mail we should know what we're doing.

If you have access to one of the pages in the "foliation" file that hasn't
been transcribed and want to try it, let me know and I'll put you down for
it.  If you want to try one and don't have one, send me a message (not the
whole list) with your address, and I'll mail you a page from the BL
microfilm, along with a copy of the Currier description page from
D'Imperio.

If the work of comparing variant transcriptions gets too onerous for me to
handle by myself, I'll ask for volunteers to do that, also.  I propose to
make as few judgements as possible myself, leaving variant readings in the
simplified cut-down minimal (enough qualifiers there?) regular expression
syntax we've discussed unless there's a clear blunder.

	Jim Gillogly

From gauss!gauss.att.com!reeds Tue Dec 10 18:09:54 EST 1991
Received: by gauss; Tue Dec 10 19:01:35 EST 1991
Received: by inet.att.com; Tue Dec 10 19:01 EST 1991
Received: from research.att.com by rand.org; Tue, 10 Dec 91 15:10:05 -0800
Message-Id: <9112102310.AA13854@rand.org>
From: reeds@gauss.att.com
Date: Tue, 10 Dec 91 18:09:54 EST
To: voynich@rand.org
Status: OR

Salvete!

Jim Gillogly suggested that I be more emphatic in announcing the
latest (noon, 10 Dec 1991) edition of my list of folios, entitled

	A Provisional Checklist of Voynich MS ``pages''

which is a valuable thing for any V-ologist to have.  It is available
by anonymous ftp from rand.org in file "pub/jim/foliation" and from me
by direct email if for any reason you are not able to ftp it.

Enjoy,

Jim Reeds
reeds@research.att.com


From gauss!math.mit.edu!jbaez Tue Dec 10 18:52:04 EST 1991
Received: by gauss; Tue Dec 10 19:18:27 EST 1991
Received: by inet.att.com; Tue Dec 10 19:18 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Tue, 10 Dec 91 15:54:49 -0800
Received: from banach (BANACH.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA21063; Tue, 10 Dec 91 18:50:50 EST
Received: by banach; Tue, 10 Dec 91 18:52:04 EST
Date: Tue, 10 Dec 91 18:52:04 EST
From: jbaez@math.mit.edu
Message-Id: <9112102352.AA12821@banach>
To: voynich@rand.org
Status: OR

Brett writes:

I would add:

6) The text is written in a line-by-line manner in language with a
positional-coded grammar (or even style).  English is an example.
We tend to place the subject at the start of the sentence.  So the
initial word has a different distribution than the entire text.
I seem to remember the style in classical Latin was to place the
verb first.  Add a grammar that encodes function in the form of the
word (as almost all languages except English do) and you would get
even stronger correlations.
----
The problem (as I see it) with this hypothesis is that the lines seem
too even in length for them to each be one sentence unless it was an
effect deliberately aimed for, as in poetry.  Nonetheless it's worth
adding this to the list.

John Baez

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Tue Dec 10 16:35 PST 1991
Received: by gauss; Tue Dec 10 19:59:09 EST 1991
Received: by inet.att.com; Tue Dec 10 19:59 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Tue, 10 Dec 91 16:37:54 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Tue, 10 Dec 91 16:42:29 PST
Date: Tue, 10 Dec 91 16:35 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: line spacing and gaps
To: voynich@rand.org
Message-Id: <D169076C221FE03879@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

It is possible that spacing has some significance.  I remember a study
of the Beowulf ms that looked at the variations in spacing between
the words and matched them against various theories of how Anglo-Saxon
verse in general, and Beiwulf in particular, might have been
read.  (R. Stevick, Suprasegmentals, Rhythm, and the Meter of Beowulf,
I think it was called, for the determined.)
 A notation probably should include information about spacing, since
after all spaces may convey information.  A truely heroic method would
be to  code each space by length so that space length mean & standard
deviation &c could be calculated.  Each space could be noted with a
space-on/length/space-off
sequence.  There might turn out to be a bi or tri modal distribution of
mean space-lengths with fairly limited variation that would enable one to
determine just what sort of space one might be looking at.

Somehow I find myself hoping that space-length coding does not have to be
carried to such, uh, lengths.

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Tue Dec 10 16:51 PST 1991
Received: by gauss; Tue Dec 10 20:48:46 EST 1991
Received: by inet.att.com; Tue Dec 10 20:48 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Tue, 10 Dec 91 16:54:50 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Tue, 10 Dec 91 16:59:34 PST
Date: Tue, 10 Dec 91 16:51 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: the manuscript was copied
To: voynich@rand.org
Message-Id: <D166E631E2DFE03879@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR


>From what I have seen of the ms pages (Brumbaugh and one or two other
facsimiles) I would also say that it gives the strong impression of
being a fair copy -- though it would take some tight speculation to
dare to suppose one could tell whether the copying was done by the
author or a copyist who knew (or did not know) what the text meant.

I'd also think (if paleographers would agree that it was a fair copy)
that it would have looked like a fair or formal copy to contemporaries:

that is, it would have looked like something prepared to be seen, rather
than looking like someone's lab notes (for example).  The equivalent
perhaps of page proofs.

This would imply that the text was either (if actually a text), a  completed
statement, in final form, or (if totally bogus) something cooked up to
look like a completed statement in final form.

Perhaps another way to look at the ms is to ask what messages its
presentation could (and couldn't) have been trying to send?

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Tue Dec 10 17:10 PST 1991
Received: by gauss; Tue Dec 10 21:00:30 EST 1991
Received: by inet.att.com; Tue Dec 10 21:00 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Tue, 10 Dec 91 17:17:04 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Tue, 10 Dec 91 17:21:48 PST
Date: Tue, 10 Dec 91 17:10 PST
From: RJB@MAX.U.WASHINGTON.EDU
To: voynich@rand.org
Message-Id: <D1642A4373BFE03879@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

When I read Laycock's book it seemed clear that despite his discretion in
expressing it his attitude was in fact skeptical.  I'd wondered what had
become of his other work -- when I was in London in '73 (I think) I
had a conversation with someone from Ambix about a linguist who collected
"artificial" languages, including one that the (only known human) speaker
claimed was Venusian (or Venerean?).  Seems that the Venusian speaking 
fellow had , in an effort to achieve a more perfect pronunciation. had
his incisors removed.

Sad to hear that there will be no more investigations into factitious 
languages by Donald Laycock.

Two more straws in favor of the Dee hypothesis.  (1) Most of Dee's
books (the ones he wrote) don't survive, but he did perhaps have a
penchant for making Books-as-objects; note what is suppsoed to be his
wedding gift to Kelley (the Tuba Veneris).  (2)  Dee was interested in
cryptography, the Enochian material shows definite signs of what I
suppose were cryptogrphoid manipulations that I suppose were characteristic
of the time -- whether they were cryptographic, or just made to suggest
that they weree, is another question entirely.

From gauss!kauri.vuw.ac.nz!sai Wed Dec 11 10:11:59 GMT 1991
Received: by gauss; Wed Dec 11 05:49:31 EST 1991
Received: by inet.att.com; Wed Dec 11 05:49 EST 1991
Received: from kauri.vuw.ac.nz by rand.org; Wed, 11 Dec 91 02:12:22 -0800
Received: by kauri.vuw.ac.nz id AA14777
  (5.65a+/IDA-1.4.2 for voynich@rand.org); Wed, 11 Dec 91 10:11:59 GMT
Date: Wed, 11 Dec 91 10:11:59 GMT
From: Simon McAuliffe <sai@kauri.vuw.ac.nz>
Message-Id: <9112111011.AA14777@kauri.vuw.ac.nz>
To: EVANS@BINAH.CC.BRANDEIS.EDU
Cc: voynich@rand.org
In-Reply-To: Ronald Hale-Evans's message of Mon, 9 Dec 1991 14:57 EDT <01GDWW0EDQ9C9S43VZ@BINAH.CC.BRANDEIS.EDU>
Subject: NSA
Status: OR

> I have a copy of Brumbaugh's *The Most Mysterious Manuscript* and frankly, I
> think it's hogslop. His methods remind me of people trying to find alphabetic
> mnemonics for telephone numbers; using the same methods, I can prove that it's
> really a vegetarian cookbook written in the late '60s. I went to Yale for my
> undergraduate studies and have visited the Beinecke more than once. I have
> never laid hands on the Ms. proper (don't ask me why; I had the privileges),
> but they gave me some photocopies that I can duplicate for people. If anyone
> has any trouble accessing the Ms. at the Beinecke, feel free to contact me. I
> believe that Yale may grant Beinecke access to alums for a fee.

Even if the book is "hogslop" I'd love to read it.  I've given up hope
of reading books like that though.  You wouldn't believe the prices of
crypto-related books over here.  They are in the order of hundreds of
dollars!  Some are as cheap as only $100 and some go over $500!  I
have to make do with what libraries have (I wonder how they can afford
them?!).

Anyway, I'm extremely interested in a copy of the Ms.  Do you still have
access to it?  Would it be possible to send me some colour negatives
(35mm or something) of it if I sent you a few $$?

> As for an introduction, I work as a user consultant at Brandeis University. I
> have been interested in cryptology since I was a child, but some of the more
> abstruse stuff like the RSA algorithm has slipped by me. To be honest, I have
> no real experience in computer analysis of text, either ciphertext or
> plaintext, but I'm willing to work with and learn from people in the Boston
> area who do. I am a contributor the the Constructed Languages mailing list,
> where I have posted information about this list. (If you would like to join that
> list, send email to conlang-request@buphy.bu.edu. I believe at least one of the
> Voynich postings from sci.lang was posted there; there should be some overlap
> of interests.) 

I too have been interested in cryptology since I was about 10 or so.
Great stuff.  I only got into it in a big way recently, including
public key cryptography (and toying with discrete logs).  I've also
been doing some cryptanalysis more recently with a reasonable degree
of success (I'm now breaking some stuff for a local company).

We're really short of people seriously interested in cryptology here
so I rarely talk to people about it.  What fields are you most
interested in?

From gauss!trl.OZ.AU!j.guy Wed Dec 11 09:40:14 EST 1991
Received: by gauss; Tue Dec 10 18:00:32 EST 1991
Received: by inet.att.com; Tue Dec 10 18:00 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Tue, 10 Dec 91 14:40:48 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA11273; Wed, 11 Dec 91 09:40:14 EST
Date: Wed, 11 Dec 91 09:40:14 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112102240.AA11273@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Dee, Enochian, and the Voynich MS
Status: OR


Dee, Enochian, and the Voynich MS

The only serious reference on Enochian, short of going back to Dee's
diaries in the British Museum, is "The Complete Enochian Dictionary - A
Dictionary of the Angelic Language as revealed to Dr John Dee and
Edward Kelley" by Donald Clarence Laycock, London: Askin Publishers,
1978. Rather, the only serious reference I knew of until today (11 Dec
91), when I read about Whitby's 1988 "John Dee's Actions With
Spirits".  

Don Laycock's book contains the complete corpus of all Enochian texts,
with interlinear translations from Dee's diaries, an Enochian-English
and English-Enochian dictionary, with a pronunciation key, and the
story of how Enochian came about. What it does not contain is Don
Laycock's personal opinion: Enochian is a fabrication by Kelley, who
was a con-artist a la Uri Geller; the phonology of Enochian is nothing
but that of Elizabethan English of the period; the vocabulary is a
mish-mash of distorted Greek, Hebrew, Latin, and names of spirits and
demons, all of which Kelley was likely to have picked up rummaging
through Dee's copious library; the grammar makes no sense, and is proof
that the language is a fabrication.

How do I know? Don Laycock and I were in the same department of the
Australian National University for 15 years, three or four doors apart,
belonged to the same poker school, and talked extensively about strange
languages, including that of the Voynich manuscript. In his "Complete
Enochian Dictionary" Don is careful to keep neutral on the matter of
the authenticity of Enochian for this understandable reason that Askin
is a publisher of occult books, and his readers would not have been
amused at reading Don's true opionions.

One interesting point: Don was convinced that Dee had owned the Voynich
manuscript, and it was Kelley who had stolen it and sold it to Rudolph
of Bohemia. His reasons to believe so were that Dee has been in Prague
with Kelley in tow at times compatible with the first mention of the
Voynich manuscript, and one remark in Dee's diary, after his return to
England, to this effect: "How strange, I cannot find my book of Soyga".
On this (to me) slim evidence Don thought that Dee's "Book of Soyga"
was in fact the Voynich manuscript, and that Kelley had swiped it.

Don Laycock died of leukemia shortly after I left the Canberran
academia for Telecom in Melbourne, much of his research on those
strange topics unpublished.


Jacques Guy  (11 Dec 91, 9:38)

From gauss!G.GP.CS.CMU.EDU!Karl.Kluge Wed Dec 11 15:39 EST 1991
Received: by gauss; Wed Dec 11 16:39:35 EST 1991
Received: by inet.att.com; Wed Dec 11 16:39 EST 1991
Received: from G.GP.CS.CMU.EDU by rand.org; Wed, 11 Dec 91 12:52:25 -0800
Date: Wed, 11 Dec 1991 15:39-EST
From: Karl.Kluge@G.GP.CS.CMU.EDU
To: voynich@rand.org
Subject: Error in D'Imperio on Brumbaugh
Message-Id: <692483988/kck@G.GP.CS.CMU.EDU>
Status: OR

While playing along at home with Brumbaugh's description of f100r, I
noticed three things:

(1) When he refers to the plant in the "lower left", he really means the
second row of plants from the top, not the bottom row. The plant on the
right of the second row is clearly the one he is calling a mandrake.
(D'Imperio noticed this also, as reflected in her transcriptions.)

(2) D'Imperio assumes he's using the same letter->digit key as the one
he gives explicitly in his book when she tries to work out the Voynich
char/digit correspondences. This is wrong, as can be seen by his
alternate readings of ulfer/alcar for one plant (i.e., a, e & u  are
equiv, f & c are equiv).

(3) Oddly enough, it seemed easier to work this through using Bennett's
notation than Currier's (I tried both)

Has anyone seen the Brumbaugh article on star names, and if so does it
contain more detail than the botany example in his book?

A transcription of my working scribbles follows.

Bennett: CT O E A Q O CT/ET O L...E/Z O CT O Q cPt...? ? C A Q
Plain:   P  A P A V A Y     J S...P   A P  E R C-US..U L F E R
                                                     A L C A R

Bennett: CT O P A U G.....E A Q CT A Q...S A M S G....Q O I? H G
Plain:   P  A L E V -US...P E P P  E R...Q U O Q -US..P A C  L -US

Bennett: ? O ? A K...E O F A L...S A K O cKt
Plain:   P J P E R...G A L A R...H E L A Y C

Vonich char (after Bennett): (A O)      (S)   (G P?) (Q E CT U?)
Plain:                       (A U E J?) (Q H) (-US)  (R V P...G? C?)


Karl

From gauss!trl.OZ.AU!j.guy Wed Dec 11 11:52:28 EST 1991
Received: by gauss; Tue Dec 10 20:35:27 EST 1991
Received: by inet.att.com; Tue Dec 10 20:35 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Tue, 10 Dec 91 16:52:49 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA11574; Wed, 11 Dec 91 11:52:28 EST
Date: Wed, 11 Dec 91 11:52:28 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112110052.AA11574@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Levitov on the Voynich MS
Status: OR


I have written a much more careful version of my original posting on sci.crypt, 
in which I examined Levitov's decipherment of the Voynich MS. Jim Gillogly has
or is about to make it ftp-able from rand.org.

From gauss!trl.OZ.AU!j.guy Wed Dec 11 12:03:03 EST 1991
Received: by gauss; Tue Dec 10 20:54:11 EST 1991
Received: by inet.att.com; Tue Dec 10 20:52 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Tue, 10 Dec 91 17:03:30 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA11591; Wed, 11 Dec 91 12:03:03 EST
Date: Wed, 11 Dec 91 12:03:03 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112110103.AA11591@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: A tool, probably useless,
Status: OR

... but you never know. I have polished up a program that, given a list
of words in different languages, tells you which words are likely to be
related to each other, and which not. By "related" I mean "fuzzily
derivable" from one another on a letter-to-letter basis. It runs only
on PCs, needs but a thimbleful of RAM, and is surprisingly fast (I once
managed to port it from our DEC KL-10 to my Kaypro II (Z80, 64k of RAM,
for those of you younger than Mathuselah), and it fitted all snugly in
30K, so imagine!). I called it COGNATE ("related" in linguists' jargon).
It comes with three sample files for testing (English, Dutch, and German).

It is, or will soon be, ftp-able from pub/jim at rand.org.

From gauss!gauss.att.com!reeds Wed Dec 11 13:00:30 EST 1991
Received: by gauss; Wed Dec 11 13:18:38 EST 1991
Received: by inet.att.com; Wed Dec 11 13:18 EST 1991
Received: from research.att.com by rand.org; Wed, 11 Dec 91 10:04:11 -0800
Message-Id: <9112111804.AA02146@rand.org>
From: reeds@gauss.att.com
Date: Wed, 11 Dec 91 13:00:30 EST
To: voynich@rand.org
Status: OR

Would it be possible for somebody living near Cambridge, Massachusetts,
to nip to the Harvard library and make a copy of Krisher's paper
for as many of us as want, for cost?  I think its in the Gordon McKay
libe, and is indexed under Voynich in the catalog, but its been years &
years, & my memory is turning to mush.  He describes (1) a graphics
transcription editor, of the sort that might be useful to us, and (2),
interesting statistical models.

Thanks in advance,

Jim Reeds.

From gauss!math.mit.edu!jbaez Wed Dec 11 14:36:28 EST 1991
Received: by gauss; Wed Dec 11 15:16:22 EST 1991
Received: by inet.att.com; Wed Dec 11 15:16 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Wed, 11 Dec 91 11:37:45 -0800
Received: from riesz (RIESZ.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA24273; Wed, 11 Dec 91 14:33:40 EST
Received: by riesz; Wed, 11 Dec 91 14:36:28 EST
Date: Wed, 11 Dec 91 14:36:28 EST
From: jbaez@math.mit.edu
Message-Id: <9112111936.AA13696@riesz>
To: voynich@rand.org
Subject: paleography
Status: OR

Nate and I spoke to an archaeologist at MIT concerning dating the Voynich
using radiocarbon or other physical means.  The difficulty with radiocarbon
would be if oil had been used to preserve the ms, or it had been fingered
too much, introducing "young" carbon.  The standard trick is to extract
the amino acids and apply radiocarbon dating to these.  The most elegant
method involves "neutron activation" with a particle accelerator (this is
how they dated the Shroud of Turin).  This would only require a tenth of a
gram, and he said that with the backing of some bigshots, esp. Yale faculty,
we might be able to convince Beinecke to hand over (directly to the lab)
a tiny piece.  He gave us the name of some radiocarbon people and someone at
Yale.  This method should be accurate to 50 years and so should be able to
distinguish between: 1) a modern forgery (e.g. by Voynich), 2) a text
written at the time of Dee, and 3) a text written at the time of Bacon.

John Baez

From gauss!cl.cam.ac.uk!Michael.Roe Wed Dec 11 20:39:21 +0000 1991
Received: by gauss; Wed Dec 11 15:57:16 EST 1991
Received: by inet.att.com; Wed Dec 11 15:57 EST 1991
Received: from swan.cl.cam.ac.uk by rand.org; Wed, 11 Dec 91 12:40:34 -0800
Received: from faeroes.cl.cam.ac.uk by swan.cl.cam.ac.uk 
          with SMTP (PP-5.7) to cl id <12691-0@swan.cl.cam.ac.uk>;
          Wed, 11 Dec 1991 20:39:33 +0000
Received: by uk.ac.cam.cl.lion (15.11/SMI-3.0DEV3) id AA04349;
          Wed, 11 Dec 91 20:39:23 gmt
Message-Id: <9112112039.AA04349@uk.ac.cam.cl.lion>
To: voynich@rand.org
Cc: Michael.Roe@cl.cam.ac.uk
Subject: C14 dating
Date: Wed, 11 Dec 91 20:39:21 +0000
From: Mike Roe <Michael.Roe@cl.cam.ac.uk>
Status: OR


In a recent message, J.Baez mentions the possibility of radiocarbon dating the
Voynich MS. A while ago, I asked a friend who works in Oxford University's
radiocarbon dating unit about this, and he was not particularly hopeful.

------- Forwarded Message

          id <aa14835>; Thu, 23 Aug 1990 14:40:40 +0000
Date: Thu, 23 AUG 90 14:39:56 +01:00
To: MRR@uk.ac.cam.cl
From: Rupert Housely 
X-Info: Research Laboratory for Archaeology and the History of Art
X-Info2: Radiocarbon Accelerator Unit
X-Info3: 6 Keble Rd, Oxford OX1 3QJ. Phone +44-865-273939
X-Info4: Fax +44-865-273932. Telex 83295 NUCLOX G.
Message-ID: </PRMD=uk.ac/ADMD= /C=gb/;gannet.cl..860:23.07.90.13.40.59>

[Deleted Text]

Regards your question about the "Voynich Manuscript".  There isn't an 
especially good technique for dating the vellum to the degree of accuracy 
which you would require.  Radiocarbon dating using an accelerator mass 
spectrometer would be able to give you a date although you would have to 
destroy approximately 30 mg of vellum in the process.  This would date 
the death of the animal from which the skin was obtained to be made into 
vellum.  It would not give the time when the ink was applied onto the vellum.
Obviously the ink could not have been applied before the animal grew it's 
skin but the ink could (theoretically) have been applied ANYTIME afterwards.
Unfortunately radiocarbon dating, being a statistical technique, has a 
standard error term which at one sigma is about +/-60 radiocarbon years.  
Because there is not a linear relationship between radiocarbon and calendar 
years it is necessary to calibrate the radiocarbon age to obtain a calendrical 
one.  The period AD c.1600-1950 is a very bad one in radiocarbon terms since 
production of 14C in the upper atmosphere kept pace with radioactive decay 
so that there is a "plateau".  This means that it is not possible to 
distinguish dates in the last few hundred years, only to say that an object 
must date to sometime within that period.  [After 1950 the atomic bomb pulse 
makes accurate dating possible].  It would be easy to check whether the vellum 
dated to the 13th century AD or to the period c.1600-1950 but it would not 
be possible to check whether it belonged to 1600-1700 or to 1912.  If it is a 
forgery it is (just) possible that the forger wrote on "old" vellum in which 
case the radiocarbon date would tell you nothing about when it was written.

------- End of Forwarded Message

Mike

From gauss!rand.org!MAILER-DAEMON Wed Dec 11 15:11:28 0800 1991
Received: by gauss; Wed Dec 11 18:11:50 EST 1991
Received: by inet.att.com; Wed Dec 11 18:11 EST 1991
Received: by rand.org; Wed, 11 Dec 91 15:11:28 -0800
Date: Wed, 11 Dec 91 15:11:28 -0800
From: Mail Delivery Subsystem <MAILER-DAEMON@rand.org>
Subject: Returned mail: Cannot send message for 1 day
Message-Id: <9112112311.AD08061@rand.org>
To: <reeds@gauss.att.com>
Status: OR

   ----- Transcript of session follows -----
421 dcs.edinburgh.ac.uk.tcp... Deferred: Connection timed out during user open with sun2.nsfnet-relay.ac.uk

   ----- Unsent message follows -----
Received: from research.att.com by rand.org; Tue, 10 Dec 91 15:10:05 -0800
Message-Id: <9112102310.AA13854@rand.org>
From: reeds@gauss.att.com
Date: Tue, 10 Dec 91 18:09:54 EST
To: voynich@rand.org

Salvete!

Jim Gillogly suggested that I be more emphatic in announcing the
latest (noon, 10 Dec 1991) edition of my list of folios, entitled

	A Provisional Checklist of Voynich MS ``pages''

which is a valuable thing for any V-ologist to have.  It is available
by anonymous ftp from rand.org in file "pub/jim/foliation" and from me
by direct email if for any reason you are not able to ftp it.

Enjoy,

Jim Reeds
reeds@research.att.com


From gauss!math.mit.edu!jbaez Wed Dec 11 21:14:28 EST 1991
Received: by gauss; Wed Dec 11 21:50:10 EST 1991
Received: by inet.att.com; Wed Dec 11 21:50 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Wed, 11 Dec 91 18:17:41 -0800
Received: from banach (BANACH.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA25679; Wed, 11 Dec 91 21:13:33 EST
Received: by banach; Wed, 11 Dec 91 21:14:28 EST
Date: Wed, 11 Dec 91 21:14:28 EST
From: jbaez@math.mit.edu
Message-Id: <9112120214.AA00179@banach>
To: voynich@rand.org
Subject: voynich stuff
Status: OR


Brett writes:

I would add:

6) The text is written in a line-by-line manner in language with a
positional-coded grammar (or even style).  English is an example.
We tend to place the subject at the start of the sentence.  So the
initial word has a different distribution than the entire text.
I seem to remember the style in classical Latin was to place the
verb first.  Add a grammar that encodes function in the form of the
word (as almost all languages except English do) and you would get
even stronger correlations.
----
The problem (as I see it) with this hypothesis is that the lines seem
too even in length for them to each be one sentence unless it was an
effect deliberately aimed for, as in poetry.  Nonetheless it's worth
adding this to the list.


Karl passes on some gloomy news concerning carbon-dating the Voynich.
I have no problem with the fact that one would actually be dating the cow
on which the Voynich was written rather than the time of writing; I seem
to recall Jacques saying that if it was newly written on velllum which
had already been used and then erased - a "palimpsest" - this might
be determined by infrared light.  (It was using infrared that they found
the signature of one of the purported owners of the text, so maybe 
it's been looked at in infrared enough already.  Nevertheless, Nate, who
has infrared viewing equipment, is itching to put it good use.)  I have
no idea what the chances are of Voynich finding lots of old unused vellum.

The business about not being able to distinguish 1600 and 1912 is a real
drag - the archaeologist I was talking to rarely concerns himself with 
anything less than a few thousand years old, so he may not have known this.
(I'll admit I'm puzzled by the cause of this "plateau".)  But it seems to 
me that distinguishing between the time of Bacon and > 1600 is worthwhile.
For example, if it turned out to be written in 1250 +/- 60, this'd be
incredibly exciting.... that is, if the cow died then.

jb

From gauss!gauss.att.com!reeds Wed Dec 11 22:29:23 EST 1991
Received: by gauss; Wed Dec 11 23:10:37 EST 1991
Received: by inet.att.com; Wed Dec 11 23:10 EST 1991
Received: from research.att.com by rand.org; Wed, 11 Dec 91 19:29:32 -0800
Message-Id: <9112120329.AA13470@rand.org>
From: reeds@gauss.att.com
Date: Wed, 11 Dec 91 22:29:23 EST
To: voynich@rand.org
Status: OR

John (or anybody likely to visit the Beinecke Libe soon) --

When you visit, before you do your FBI crime lab & shroud of
Turin number, please check all the fold out sheets.  Note how
the creases lie, and whether they are ridges or valleys, and in
which order the folds must be made to fold up the sheet again.
Think of it as computational origami: I want the program and
the data.  Fold up a sheet of note book paper then & there 
the same way, note on the sheet which edge goes in the spine,
and write on the notebook page sufficient info for me to tell
what image goes where.  (A brief sketch, say, or a few words
of incipit.)

Thanks in advance!

Jim Reeds


From gauss!im.lcs.mit.edu!nate Wed Dec 11 23:33:47 EST 1991
Received: by gauss; Wed Dec 11 23:50:15 EST 1991
Received: by inet.att.com; Wed Dec 11 23:50 EST 1991
Received: from IM.LCS.MIT.EDU by rand.org; Wed, 11 Dec 91 20:34:01 -0800
Received: from im12 (IM12.LCS.MIT.EDU) by im (4.1/TOC-1.1S) 
	id AA15388; Wed, 11 Dec 91 23:33:48 EST
From: nate@im.lcs.mit.edu (Nate Osgood)
Received: by im12 (4.1/TOC-1.1C) 
	id AA01310; Wed, 11 Dec 91 23:33:47 EST
Date: Wed, 11 Dec 91 23:33:47 EST
Message-Id: <9112120433.AA01310@im12>
To: voynich@rand.org
Subject: Dating techniques
Status: OR

Thanks to Michael Roe for his informative message.  Sounds like C14
dating is not a viable method for discriminating between more recent
dates of possible creation (assuming that the age of the vellum is an
accurate reflection of date of writing), although I would still regard
such a technique as desirable for distinguishing between the 13th
century and later document.   30mg of (miminally contaminated) vellum
is probably extractable without  damaging the document in any
significant manner (yet another reason that underscores the importance
of seeing the document first hand).

Given the problems of applying C14 dating to recent material, I wonder
about the applicability of amino acid racemization dating techniques.
These are starting to come into serious use in paleoanthropology,
although some difficulties may remain to be ironed out.  I know AAR
tends to be applicable over a significantly longer period of time than
conventional radiocarbon dating (which only brings one back to 40K
BP), and it may be that the resolution would be too coarse to
distinguish between e.g. a 16th century and 20th century piece of
vellum.)  Nonetheless, the technique deserves some research as a
possible means of addressing the dating problem.  It should be
stressed that this technique, like C14 dating,  only addresses the
question as to when the animal lived.  (As far as I understand the
discussion reported  to me prior to my joining this list, the
assumption that the vellum is the same age as the documents seems
relatively safe -- I guess obtaining clean vellum would otherwise be
too difficult).    I'll do a bit of research into AAR and see if it
looks applicable.  (There may be other techniques that would fit the
bill as well, although I can't think of any off hand).

Nate






From gauss!im.lcs.mit.edu!nate Thu Dec 12 00:02:02 EST 1991
Received: by gauss; Thu Dec 12 00:34:27 EST 1991
Received: by inet.att.com; Thu Dec 12 00:34 EST 1991
Received: from IM.LCS.MIT.EDU by rand.org; Wed, 11 Dec 91 21:02:14 -0800
Received: from im12 (IM12.LCS.MIT.EDU) by im (4.1/TOC-1.1S) 
	id AA15647; Thu, 12 Dec 91 00:02:03 EST
From: nate@im.lcs.mit.edu (Nate Osgood)
Received: by im12 (4.1/TOC-1.1C) 
	id AA01316; Thu, 12 Dec 91 00:02:02 EST
Date: Thu, 12 Dec 91 00:02:02 EST
Message-Id: <9112120502.AA01316@im12>
To: voynich@rand.org
Subject: Random ideas
Status: OR

Please excuse this posting if it repeats a lot of material discussed
prior to my emergence on the list.  I've been particularly interested
in trying to find PHYSICAL techniques which could help find clues to
the origin of the  manuscript.  As as suggested in my last message, I
think there remains a possible role for dating techniques in this
problem.  Nonetheless, I think there may be a variety of other cues
that might be able to help as well.  Unfortunately, I think that many
of them are linked to questionable assumptions, while others would
require substantial knowledge of very specialized areas of study.   I
thought I'd post a few of my thoughts on possible means of attack,
just to see if anyone else out there might be able to provide  more
definitive judgements as to their applicability:

Analysis of inks:   It seems that given an adequate sample of the
voynich ink, it might be possible to place it in a particular time
period (or even place).  I heard somewhere that some of the ink looks
suspiciously similar to modern varieties.   Such affinities would seem
to be rather easy to check, given the appropriate removal and  chemical
analysis  techniques, coupled with some knowledge of how ink
technology changed over time.  It seems reasonable to assume that
while Voynich (or any other forger) may have conceivably been able to
get their hands on  ancient vellum, it seems unlikely  they would have
been able to obtain ancient, usable ink.  (It might also be profitable
to see whether PEN technology changed much between the different
periods -- would 13th century pens  have been really crappy compared
to e.g. 16th century pens, and left characteristic signatures?)

Analysis of book construction (both overall method and associated materials
such as glue):  While any careful forger might have been able to
simulate the book construction technology of a given period, it seems
unlikely that e.g. the glue technology could be well simulated.  If the
binding style is more modern (e.g. 17th century), it would at least
provide SOME evidence to suggest that an earlier century was not
responsible for the book.  (Of course, the book could easily have been
rebound, although presumably not without leaving some tell-tale  signs
of THIS activity)

Research into medieval/renaissance literary practices.  This sounds
pretty vague, and it is.  I guess I'm thinking that it MIGHT be useful
to know somethink along the lines of WHO would have been able to write
a book  of this sort, and WHO would actually have done the writing.
Would Roger Bacon have actually scribed a book himself, or hired a
scribe to do it.  If the former, could we compare handwriting to other
books?  If the latter, might not we account for different handwritings
in the manuscript by the presence of different scribes?  Would a
scribe have been hired by an alchemist to write secrets in some
bizarre script, or would the achemist have done the job themselves?
These questions sound somewhat incidental to the whole mystery, but I
wouldn't be terribly surprised if a good understanding of the
environments/practices that would have surrounded the book at
different time periods could be used as some sort of clues in the
whole mystery.

Faded writing: I'm sure this has been belabored many times over, but
there are many observational techniques that would allow one to
perceive many contrasts not visible in the visual region of the
spectrum that would relate to "faded" writing.  (Chemical techniques
exist as well, although they are less desirable for the obvious
reasons of irreversability).  Infrared and ultraviolet observation can
been particularly useful in this regard, and has been applied to good
effect in in archaeology, law enforcement and art restoration.
Taking a few pictures of the manuscript under infrared film and
observing the manuscript through an infrared image tube could yield
some very interesting information (particularly if there is indeed a
faded "key" page :-) )  

In any case, these were a few of the possible avenues of attack I
envision as possibly remaining open.   Any ideas as to how realistic
they might be?

From gauss!trl.OZ.AU!j.guy Thu Dec 12 09:26:29 EST 1991
Received: by gauss; Wed Dec 11 17:53:44 EST 1991
Received: by inet.att.com; Wed Dec 11 17:53 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Wed, 11 Dec 91 14:26:40 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA13322; Thu, 12 Dec 91 09:26:29 EST
Date: Thu, 12 Dec 91 09:26:29 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112112226.AA13322@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: A utility (*USEFUL*)
Status: OR


I have sent to Jim Gillogly a biggish program I wrote last year -- or
is it the year before last -- to compute the entropy of texts to any
order, generate random text, and produce on-screen concordances. Like
everything I write, it runs only on PCs. The file I sent will unzip
to:


MONKEY.EXE  (58,272 bytes)
MONKEY.DOC  (22,441 bytes)

MONKEY is a fairly souped-up implementation of the algorithm describe
in my article in Glottometrika: "Fast high-order monkeys and a fast
algorithm for calculating high-order character entropies",
Glottometrika 12, Bochum: Universitaetsverlag Dr N Brockmeyer,
1990:125-130.

However, I suspect an error in the code that shows up infrequently. I
have observed that, sometimes, the calculated value of the x-th order
entropy is grossly wrong. For instance, you may get a 3rd-order
character entropy of 2.3, 5th-order of 0.8, with in-between a 4th-order
value of 10.5! When you have MONKEY read in the same file again, the
error usually disappears (yes, I know what you're thinking: but it's
5000 lines of OOP Pascal I have to wade through, built up by adding
feature upon feature, piece-meal. Better rewrite the whole thing from
scratch).

I have time and again gone to the trouble of computing the entropy by
hand, only to find MONKEY giving me the right answers (those I had
calculated) time and again. Perverse. I have been writing a new version
of MONKEY from scratch, with a better interface (accepts a mouse), and
no limitation to the size of the corpus (as long as both corpus and
index will fit in RAM), but it's been an on and off job, mostly off. On
the principle that an imperfect tool is better than no tool at all,
well, there it is, or soon will be. Look for monkey.zip in pub/jim at
rand.org.

From gauss!im.lcs.mit.edu!nate Thu Dec 12 10:33:24 EST 1991
Received: by gauss; Thu Dec 12 11:11:14 EST 1991
Received: by inet.att.com; Thu Dec 12 11:11 EST 1991
Received: from IM.LCS.MIT.EDU by rand.org; Thu, 12 Dec 91 07:34:25 -0800
Received: from im7 (IM7.LCS.MIT.EDU) by im (4.1/TOC-1.1S) 
	id AA21911; Thu, 12 Dec 91 10:33:26 EST
From: nate@im.lcs.mit.edu (Nate Osgood)
Received: by im7 (4.1/TOC-1.1C) 
	id AA02899; Thu, 12 Dec 91 10:33:24 EST
Date: Thu, 12 Dec 91 10:33:24 EST
Message-Id: <9112121533.AA02899@im7>
To: voynich@rand.org
Subject: Dee's writing
Status: OR

If these things can be quantified, it might also be interesting to
compare Dee's handwriting (assuming it WAS he who physically wrote his
documents) in other, existing books to the writing in the Voynich.

Do any illustrations exist in any other of Dee's works?   The
idiosyncratic nature of those illustrations might be recognizable in
other books as well.

Nate


From gauss!G.GP.CS.CMU.EDU!Karl.Kluge Thu Dec 12 16:06 EST 1991
Received: by gauss; Thu Dec 12 16:43:36 EST 1991
Received: by inet.att.com; Thu Dec 12 16:43 EST 1991
Received: from G.GP.CS.CMU.EDU by rand.org; Thu, 12 Dec 91 13:17:15 -0800
Date: Thu, 12 Dec 1991 16:06-EST
From: Karl.Kluge@G.GP.CS.CMU.EDU
To: voynich@rand.org
Subject: Bennett <--> Currier correspondence
Message-Id: <692572014/kck@G.GP.CS.CMU.EDU>
Status: OR

Someone requested the correspondence to use with my post on Brunbaugh's
herbal labels.

Bennett: A I L M N O C E  T P H F K Q U V Y Z D S G CT ET cPt eHt cFt
Currier: A I E M N O C 2? 1 B P V F R D * * 2 4 8 9 S  Z  W   Q?  Y

Bennett "v" & "y" are as per the First Study group system -- they look
like those Roman letters upside down. 

From gauss!gauss.att.com!reeds Thu Dec 12 11:34:38 EST 1991
Received: by gauss; Thu Dec 12 11:56:26 EST 1991
Received: by inet.att.com; Thu Dec 12 11:56 EST 1991
Received: from [192.20.225.2] by rand.org; Thu, 12 Dec 91 08:36:00 -0800
Message-Id: <9112121636.AA26397@rand.org>
From: reeds@gauss.att.com
Date: Thu, 12 Dec 91 11:34:38 EST
To: voynich@rand.org
Status: OR

Nate writes:

> Do any illustrations exist in any other of Dee's works?

Yes.  Check out any of the recent books about Dee.  It is not
at all impossible that he was the author.  But consider: (1) Dee
was an honest man, and (2) Dee might have been the victim not
author of hoax, so the hoaxer prepares it so it will appeal to
Dee's known style, or (3) Dee's style is shared by other late 
16-th century astrology hackers.  Have you read Francis Yeats's
books, which give a general survey of the ambient Neoplatonic 
circle of ideas which would have been shared (known, if not adhered
to) by hoaxer and hoaxee at the time?

Sorry I don't have any particular books abot Dee to recommend, but
but he's a fairly popular target (and has been, ever since Ben Johnson),
so your favorite library should turn something up.

Jim Reeds

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Thu Dec 12 08:52 PST 1991
Received: by gauss; Thu Dec 12 12:41:06 EST 1991
Received: by inet.att.com; Thu Dec 12 12:41 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Thu, 12 Dec 91 08:53:49 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Thu, 12 Dec 91 08:58:31 PST
Date: Thu, 12 Dec 91 08:52 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: carbon dating cows
To: voynich@rand.org
Message-Id: <D017689AC3FF0020E9@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

This would suggest that the text was not written, but tattooed.

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Thu Dec 12 08:57 PST 1991
Received: by gauss; Thu Dec 12 12:52:24 EST 1991
Received: by inet.att.com; Thu Dec 12 12:52 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Thu, 12 Dec 91 08:59:40 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Thu, 12 Dec 91 09:04:22 PST
Date: Thu, 12 Dec 91 08:57 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: computational origami and paleography
To: voynich@rand.org
Message-Id: <D016AFD177BF0020E9@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR


Yes:  it would be interesting to try to surmise whether the text was
written into a bound book or whether the sheets were codexicated (!) after
they were written.

A problem by the way that may be old hat by now is that even in fair
copies manuscripts are often full of abbreviations, and one would have
to take that possibility into account.

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Thu Dec 12 09:18 PST 1991
Received: by gauss; Thu Dec 12 13:01:23 EST 1991
Received: by inet.att.com; Thu Dec 12 13:01 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Thu, 12 Dec 91 09:22:23 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Thu, 12 Dec 91 09:27:06 PST
Date: Thu, 12 Dec 91 09:18 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Who could have done it?
To: voynich@rand.org
Message-Id: <D013D9723C1F0020E9@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

Nate's query about who could have written the actual text:  of course
at any point a copyist could have been hired to produce a facsimile
of an unreadable curiosity.  But educated people did in fact do a lot
of transcribing and fair copying of texts themselves.  To take a
recurrent figure, John Dee prepared both immediate transcripts and later
very neat fair copies of his seances with EK (the BL has bales of the
stuff); he also had another hand for writing formal documents (see his
letters of petition for example).

Another question:  physical properties of the text.  Many bookbinders
techniques have remained fairly unchanged over the centuries, despite
the ravages of the degenerate 18th and 19th centuries.  Medieval 
manuscripts at their best are superb pieces of engineering that will
probably be around long after most modern books have become dust -- 
and yet the materials were simple and easily available.  In other words,
the use of old techniques or materials would prove little, since people
still bind books using medieval stitching and adhesives.  A 19th or  early
20th century forger might well, however, have used glue or paste or
cord with anachronistic physical or chemical properties.  It might also
be worth looking at the sizing if any on the ms pages.

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Thu Dec 12 09:31 PST 1991
Received: by gauss; Thu Dec 12 13:19:16 EST 1991
Received: by inet.att.com; Thu Dec 12 13:19 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Thu, 12 Dec 91 09:33:21 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Thu, 12 Dec 91 09:37:59 PST
Date: Thu, 12 Dec 91 09:31 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Dee's writing
To: voynich@rand.org
Message-Id: <D011F4BC3C7F0020E9@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

There are sketches and illustrations in some of the Enochian material,
and in the _Tuba Veneris_ (though there seems to be some debate about
whether that is genuine, and I haven't had any first-hand access to the
debate so I don't know what the issues are).

Much of the Dee material (at the BL, at any rate) has been microfilmed
or photographically copied, so negatives exist and it would probably
be not too hard to do some comparisons, if one knew what pages one 
wanted from what mss!
    
The _Tuba Veneris_, last time I heard, was held at the Warburg, and they
seemed to feel that some debate about ownership and perhaps about 
scholarly priority (as in the Dead Sea scrolls) prevented them from
producing a microfilmed copy.  (My news on this is very old news
indeed, and if anyone has anything more recent, I'd be pleased to hear
it.)

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Thu Dec 12 09:42 PST 1991
Received: by gauss; Thu Dec 12 13:23:05 EST 1991
Received: by inet.att.com; Thu Dec 12 13:23 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Thu, 12 Dec 91 09:47:52 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Thu, 12 Dec 91 09:51:58 PST
Date: Thu, 12 Dec 91 09:42 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Elizabethan astrological and alchemical hackers
To: voynich@rand.org
Message-Id: <D0107425F39F0020E9@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

Of course it *is* possible that the ms was not intended as a hoax, no
matter how it may have been used.

Or, if intended as a hoax, it could (on the cryptographic hypothesis)
have been intended as a demontration hoax.

I think it is too early toi use such consiedrations to close off
possible approaches -- they are most useful as methods of suggesting
more divergent approaches.  Too little is yet clearly established.
If physical analysis made it clear that the ms as an object could not
have existed before the 19th century (for example), *that* would
rule out any need to hyopthesize about Dee.  (too many p's in
"hypothesize"!).  (too many typos entirely; oh well)

It's true that ability to write a fair hand and sketch presentably
was not a rarity among the literate of the period; nor was an interest
in secret writing, and so on.  Dee is not the only candidate -- just
the only one well known enough to be the standard reference candidate.

From gauss!math.mit.edu!jbaez Thu Dec 12 13:10:12 EST 1991
Received: by gauss; Thu Dec 12 13:58:56 EST 1991
Received: by inet.att.com; Thu Dec 12 13:58 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Thu, 12 Dec 91 10:13:24 -0800
Received: from banach (BANACH.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA28071; Thu, 12 Dec 91 13:09:16 EST
Received: by banach; Thu, 12 Dec 91 13:10:12 EST
Date: Thu, 12 Dec 91 13:10:12 EST
From: jbaez@math.mit.edu
Message-Id: <9112121810.AA00975@banach>
To: voynich@rand.org
Subject: examining the ms
Status: OR

As I intend to take a look at the ms sometime soon, I want to make that
visit as productive as possible.  Of course, having a decent paleographer
along would be the best solution ---- does anybody know one near Yale who
could be enticed into coming along?  Anyway, I'm curious about what this
remark by the mysterious "RJB" means:

 It might also be worth looking at the sizing if any on the ms pages.

John Baez

From gauss!rand.org!MAILER-DAEMON Thu Dec 12 10:36:06 0800 1991
Received: by gauss; Thu Dec 12 13:36:23 EST 1991
Received: by inet.att.com; Thu Dec 12 13:36 EST 1991
Received: by rand.org; Thu, 12 Dec 91 10:36:06 -0800
Date: Thu, 12 Dec 91 10:36:06 -0800
From: Mail Delivery Subsystem <MAILER-DAEMON@rand.org>
Subject: Returned mail: Cannot send message for 1 day
Message-Id: <9112121836.AA27682@rand.org>
To: <reeds@gauss.att.com>
Status: OR

   ----- Transcript of session follows -----
421 dcs.edinburgh.ac.uk.tcp... Deferred: Connection timed out during user open with sun2.nsfnet-relay.ac.uk

   ----- Unsent message follows -----
Received: from research.att.com by rand.org; Wed, 11 Dec 91 10:04:11 -0800
Message-Id: <9112111804.AA02146@rand.org>
From: reeds@gauss.att.com
Date: Wed, 11 Dec 91 13:00:30 EST
To: voynich@rand.org

Would it be possible for somebody living near Cambridge, Massachusetts,
to nip to the Harvard library and make a copy of Krisher's paper
for as many of us as want, for cost?  I think its in the Gordon McKay
libe, and is indexed under Voynich in the catalog, but its been years &
years, & my memory is turning to mush.  He describes (1) a graphics
transcription editor, of the sort that might be useful to us, and (2),
interesting statistical models.

Thanks in advance,

Jim Reeds.

From gauss!gauss.att.com!reeds Thu Dec 12 13:39:21 EST 1991
Received: by gauss; Thu Dec 12 14:37:43 EST 1991
Received: by inet.att.com; Thu Dec 12 14:37 EST 1991
Received: from research.att.com by rand.org; Thu, 12 Dec 91 10:39:38 -0800
Message-Id: <9112121839.AA29330@rand.org>
From: reeds@gauss.att.com
Date: Thu, 12 Dec 91 13:39:21 EST
To: voynich@rand.org
Status: OR

Salvete!

A much corrected list of folios and images in rand.org:pub/jim/foliations,
thanks to data supplied by Karl Kluge.  (And thanks to technology change:
I went over to using 3 by 5 cards, which are much more efficient for this
task than anything on the computer!)

Here is a breakdown of the 232 ``pages'' I think exist:
					
					In D'Imp's transc.	Not in transc.
Image Published					16			18
Image not published but on Jim G's BM film	103			0
Image not published nor on BM film.		12			83

Totals						131			101

We have transcribed some (maybe half a dozen) pages so far, but I regard
them as unchecked drafts or practice runs.  Of course the Beinecke
film covers the whole MS, but we don't have a copy of that yet.  And
some of the forms of ``publication'' listed above are pretty obscure,
as are some of the images.

Assuming that we transcribe each of D'Imps pages once, and each of the
others twice, the bare minimum needed to get a check on everything, we
have 333 pages to go: halfway to perdition!

Merry Michiton, and a happy Oladabas Day to you!

Jim Reeds

From gauss!math.mit.edu!jbaez Thu Dec 12 13:43:14 EST 1991
Received: by gauss; Thu Dec 12 14:50:06 EST 1991
Received: by inet.att.com; Thu Dec 12 14:49 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Thu, 12 Dec 91 10:46:25 -0800
Received: from banach (BANACH.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA28219; Thu, 12 Dec 91 13:42:17 EST
Received: by banach; Thu, 12 Dec 91 13:43:14 EST
Date: Thu, 12 Dec 91 13:43:14 EST
From: jbaez@math.mit.edu
Message-Id: <9112121843.AA00983@banach>
To: voynich@rand.org
Subject: Hypotheses versus Data
Status: OR

While hypothesizing about who could have written the Voynich and why is:
1) irresistible, and 2) helpful for coming up with research directions,
it should be clear that it will not solve the problem.  I hesitate to point
this out since to some of you it's utterly clear, but perhaps to others it
is not.   Everyone who wants to try their hand at such hypotheses MUST read
D'Imperio's book, because it's the authoritative work on such hypotheses and
the evidence.  One comes away from this book with a clear sense that what's neeed is more solid DATA.  Reed, Guy and Gillogly are on the right track here itting their collective nose to the grindstone to squeeze more data from the
text.  I promise to do some transcribing this weekend when I'll be less 
distracted by this darn paper I'm trying to finish up.   Physical analyses
of the text, as Nate proposes, could also be very helpful.

Have I got this right: that only Ronald Hale-Evans has a full photocopy of the
ms right now?  I'm getting sick of waiting to hear from Beinecke and am 
itching for a photocopy.  How much would it cost?

jb

From gauss!math.mit.edu!jbaez Thu Dec 12 13:47:28 EST 1991
Received: by gauss; Thu Dec 12 14:40:36 EST 1991
Received: by inet.att.com; Thu Dec 12 14:40 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Thu, 12 Dec 91 10:50:38 -0800
Received: from banach (BANACH.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA28244; Thu, 12 Dec 91 13:46:31 EST
Received: by banach; Thu, 12 Dec 91 13:47:28 EST
Date: Thu, 12 Dec 91 13:47:28 EST
From: jbaez@math.mit.edu
Message-Id: <9112121847.AA00991@banach>
To: voynich@rand.org
Subject: photocopies
Status: OR

By the way, if I get Hale-Evans photocopies (and we're in the same town so
it should be logistically simple), I can send copies of copies of copies
(that's what they'll be, alas) to some of you FOR FREE, since I can justify
this to the department as academic business.

jb

From gauss!trl.OZ.AU!j.guy Thu Dec 12 14:11:57 EST 1991
Received: by gauss; Wed Dec 11 22:31:01 EST 1991
Received: by inet.att.com; Wed Dec 11 22:30 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Wed, 11 Dec 91 19:12:45 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA13686; Thu, 12 Dec 91 14:11:57 EST
Date: Thu, 12 Dec 91 14:11:57 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112120311.AA13686@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Carbon dating cows
Status: OR


John Baez gets all het up:

"For example, if it turned out to be written in 1250 +/- 60, this'd be
incredibly exciting.... that is, if the cow died then."

Hold your horses there! Indeed we must not overlook the possibility
that the Voynich was written on *LIVE* cows, for there is a story that
some eccentric Briton once wrote a cheque on a cow, and British law
being what it was, the cheque was valid, and... Now I remember quite
clearly that it was not Roger Bacon, nor Francis Bacon (a.k.a. William
Shakespeare) who did it. Could it have been John Dee? Ho, hum, living
in Australia where the seasons are reversed, you sort of tend to
confuse Xmas and April 1st, you know.

From gauss!uunet.UU.NET!cbmvax!snark!cowan Thu Dec 12 15:43:15 EST 1991
Received: by gauss; Thu Dec 12 17:05:57 EST 1991
Received: by inet.att.com; Thu Dec 12 17:05 EST 1991
Received: from relay2.UU.NET by rand.org; Thu, 12 Dec 91 13:04:10 -0800
Received: from uunet.uu.net (via LOCALHOST.UU.NET) by relay2.UU.NET with SMTP 
	(5.61/UUNET-internet-primary) id AA05448; Thu, 12 Dec 91 16:04:14 -0500
Received: from cbmvax.UUCP by uunet.uu.net with UUCP/RMAIL
	(queueing-rmail) id 160306.9107; Thu, 12 Dec 1991 16:03:06 EST
Received: by cbmvax.cbm.commodore.com (5.57/UUCP-Project/Commodore 2/8/91)
	id AA03959; Thu, 12 Dec 91 16:01:02 EST
Received: by snark.thyrsus.com (/\==/\ Smail3.1.21.1 #21.19)
	id <m0krxFU-00007PC@snark.thyrsus.com>; Thu, 12 Dec 91 15:43 EST
Message-Id: <m0krxFU-00007PC@snark.thyrsus.com>
From: cbmvax!snark.thyrsus.com!cowan@uunet.UU.NET (John Cowan)
Subject: Re: Carbon dating cows
To: voynich@rand.org (Voynich List)
Date: Thu, 12 Dec 91 15:43:15 EST
In-Reply-To: <9112120311.AA13686@medici.trl.OZ.AU>; from "Jacques Guy" at Dec 12, 91 2:11 pm
X-Mailer: ELM [version 2.3 PL11]
Status: OR

Jacques Guy writes:

> [T]here is a story that
> some eccentric Briton once wrote a cheque on a cow, and British law
> being what it was, the cheque was valid, and...

Sorry, but that episode is strictly fictional.  "The Negotiable Cow" appeared
in >Punch< and was clearly labelled as fiction; the alleged perpetrator
is the hero of several other law cases by the same (actual) author.

Notwithstanding, a cheque written on a cow would indeed be negotiable; nothing
says that the assignee (the person to whom the cheque is made out) needs to
>accept< such a thing, however.  "Assignation is not payment."

-- 
cowan@snark.thyrsus.com		...!uunet!cbmvax!snark!cowan
		e'osai ko sarji la lojban

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Thu Dec 12 13:28 PST 1991
Received: by gauss; Thu Dec 12 17:22:04 EST 1991
Received: by inet.att.com; Thu Dec 12 17:21 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Thu, 12 Dec 91 13:31:08 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Thu, 12 Dec 91 13:35:52 PST
Date: Thu, 12 Dec 91 13:28 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: sizing
To: voynich@rand.org
Message-Id: <CFF0D86B1A3F003492@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

I can't remember off-hand about velum, but paper (in the West, at any
rate:  different writing technology in China and Japan) is almost
always "sized" so that the ink won't blot.  (Blotting paper on the other
hand...)  Different kinds of materials have been used for sizing; nowadays
in marbling paper for example one often uses an alum solution.  (Well,
actually, that remark may be out of date.  But it gives you the general
idea.)

It would be interesting to look at the pages to see if they had been sized or
otherwise treated and, if so, with what.  And also to look at the 
illustrations, especially when colors are used, to see what the pigments
might have been, and what the binder might have been.  Gold was often
attached with glaire (an eggwhite adhesive) for example.

A forger *might* have used preservatives not available at the time, in
any of the adhesives or other preparations.  I have used thymol and oil of
cloves in wheat paste -- there is a faint but characteristic odor.  I haven't
come back though after 500 years to see how well it lasts...

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Thu Dec 12 13:51 PST 1991
Received: by gauss; Thu Dec 12 17:25:41 EST 1991
Received: by inet.att.com; Thu Dec 12 17:25 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Thu, 12 Dec 91 13:55:07 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Thu, 12 Dec 91 13:59:51 PST
Date: Thu, 12 Dec 91 13:51 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: grindstones
To: voynich@rand.org
Message-Id: <CFED94F22E9F003492@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR


You are of course quite right; my only justification for the needless
multiplication of hypotheses is that one can think of new kinds of data to
look for.

I should read D'Imperio's book.  The library here doesn't have a copy; our
Government Documents librarians insist that they did not find it listed
as part of any gov't/NSA technical report series.  I am going to
try interlibrary loan -- but do not look for quick results.  I have also
dropped a note to the NSA, addressed in the vaguest of terms to Fort
Mead MD (on the grounds that hunting up a more specific address might
draw more than bibliographic attention), and again have not managed to
convince myself that I'll get a quick response.  It doesn't seem to be in
Books In Print.

A large class of government publications are not copyrighted:  is this
one?

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Thu Dec 12 14:02 PST 1991
Received: by gauss; Thu Dec 12 19:00:42 EST 1991
Received: by inet.att.com; Thu Dec 12 19:00 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Thu, 12 Dec 91 14:33:23 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Thu, 12 Dec 91 14:16:01 PST
Date: Thu, 12 Dec 91 14:02 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: film and photocopies
To: voynich@rand.org
Message-Id: <CFEC0ECDC73F003492@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR


If the Beinecke has only a microfilm, and the BL has the actual ms, as I
understand the BL's way of doing things, the first person to have
a ms filmed pays for the negtive and a positive.  The negative remains
with the BL, and scholars can order positives from that negative at
a somewhat more bearable price.

If that is the situation, it should be relatively easy (though not
necessarily fast), and relatively inexpensive, to get a microfilm
print directly from the Department of Manuscripts at the BL.

Does the BL have the ms?  Is the Beinecke film a positive from the
BL's negative?  I know that the BL would much rather have people
order films than have them actually pawing their mss:  I don't
even think one needs to have a reader's ticket to the manuscript room
to be able to order prints or microfilms -- just some reasonably
respectable academic connection.  It's been more than 6 years though
since I was last there and things -- there or in my memory-- may well
have changed.

From gauss!math.mit.edu!jbaez Thu Dec 12 17:29:50 EST 1991
Received: by gauss; Thu Dec 12 18:49:26 EST 1991
Received: by inet.att.com; Thu Dec 12 18:49 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Thu, 12 Dec 91 14:33:01 -0800
Received: from banach (BANACH.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA29318; Thu, 12 Dec 91 17:28:53 EST
Received: by banach; Thu, 12 Dec 91 17:29:50 EST
Date: Thu, 12 Dec 91 17:29:50 EST
From: jbaez@math.mit.edu
Message-Id: <9112122229.AA01404@banach>
To: voynich@rand.org
Subject: D'Imperio
Status: OR

I don't have D'Imperio on me so I don't recall if it's copyrighted.
It is definitely published by the NSA.  Since the government has been
known to claim there is No Such Agency I'm not too surprised at your
troubles in finding it.  :-)  

For those who are 1) utterly desperate and 2) planning on doing a lot of
transcribing, I might be persuaded to send you big hunks of the D'Imperio.

jb

From gauss!math.mit.edu!jbaez Fri Dec 13 11:31:58 EST 1991
Received: by gauss; Fri Dec 13 12:21:23 EST 1991
Received: by inet.att.com; Fri Dec 13 12:21 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Fri, 13 Dec 91 08:33:17 -0800
Received: from riesz (RIESZ.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA01685; Fri, 13 Dec 91 11:29:01 EST
Received: by riesz; Fri, 13 Dec 91 11:31:58 EST
Date: Fri, 13 Dec 91 11:31:58 EST
From: jbaez@math.mit.edu
Message-Id: <9112131631.AA03886@riesz>
To: voynich@rand.org
Subject: miscellaneous news
Status: OR

I got a copy of the manuscript today - no, not the Voynich, the galley
proofs of the book on quantum field theory of which I'm a coauthor.  This
will make a big dent in my spare time for a while, but then will be gone
forever, thank god. 

The D'Imperio book has NO COPYRIGHT on it so I will be unabashed about
xeroxing it and mailing people copies.

D'Imperio mentions that Tiltman's findings are the most worthwhile work
to have been done on the Voynich.  Unfortunately they are unpublished.
She says she'll be glad to pass on stuff to serious scholars.  Any chance that
someone (e.g. Jim Gillogly :-) could track her down and get that stuff?  

I've been getting "failed mail" bounces of my postings from 
 raven.eklektix.com.  I presume my postings are getting through to the
rest of you; if someone associated with that site could fix something that'd
be nice...

Off to Bard College, New York, and proofreading my ms and transcribing X's
ms.  

jb

From gauss!gauss.att.com!reeds Fri Dec 13 12:43:30 EST 1991
Received: by gauss; Fri Dec 13 12:57:04 EST 1991
Received: by inet.att.com; Fri Dec 13 12:56 EST 1991
Received: from research.att.com by rand.org; Fri, 13 Dec 91 09:43:42 -0800
Message-Id: <9112131743.AA21762@rand.org>
From: reeds@gauss.att.com
Date: Fri, 13 Dec 91 12:43:30 EST
To: voynich@rand.org
Status: OR

Gang,

I feel foolish spending so much time on procedural issues, instead
of actually transcribing, but:  As I've been transcribing I've kept
track of a few problems that have come up, and have collected them
into a transcriber's guidelines, a draft of which follows.  Any
contributions?

Jim Reeds.


The Prescott Currier alphabet has 36 symbols, but it is not always 
easy to see how to turn a bit of V text into PC letters.
These notes contain accumulated wisdom:

1. The appearances of J and 7 are so close that nobody I know can tell
them apart.  Maybe J has a bigger swashier tail? If you spot any page 
of D'Imperio's transcription showing clear cut distinguishable examples 
of both J and 7 let us know.

2. Sometimes the tall 2 legged letters like P and F have their 
feet planted in different parts of a word, or even in different words; 
the pedestal versions (Q and so on) have separate ``S'' shoes, one
for each foot.  In this case (1) transcribe it as two appearances of
the letter (located where the feet land), but (2) put a note in the 
comments section.

3. The C is so small and occurs in combinations like CC and CCC which
look more like ordinary "u" or "omega", "ui" or "iu".  Maybe the Currier
alphabet should be extended by adding "u" and "o" for CC and CCC joined
together to look like single lower-case "u" and "omega"? Go ahead and 
use these symbols if you want, but put flag it with a comment.  (Be
sure to keyboard o and O distinctly, and u and U!)

4. The 2 and R symbols are usually easy to tell apart: the 2 is a C (a
tiny little C) with a plume, and the R is a straight stroke (NW to SE)
with a plume.  The R's stroke is somewhat longer than the little C in the 2.
At the bottom of the plume, the R has a cusp and the 2 has a smooth spline.
If still in doubt use [2R].

5. Word spaces are tough to call.  Use the general spacing on the page
as a guide.  Use , for a doubtful word space.  If an illustration
obtrudes put in a "-" and a comment.  If you are in doubt about whether
it obtrudes, put in at least a comment.

6. Sometimes you see a character like an upsidedown  "u", or like one
of the roller-coaster humps of a lower-case "m".  It might be part of
a larger Currier letter: 0, 1, 3, H, K, L, M, N, T, or U!

From gauss!rand.org!jim%mycroft Fri Dec 13 11:10:07 PST 1991
Received: by gauss; Fri Dec 13 14:38:04 EST 1991
Received: by inet.att.com; Fri Dec 13 14:37 EST 1991
Received: from mycroft.rand.org by rand.org; Fri, 13 Dec 91 11:10:10 -0800
Received: from localhost by mycroft.rand.org; Fri, 13 Dec 91 11:10:07 PST
Message-Id: <9112131910.AA12386@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: miscellaneous news [D'Imperiod, Tiltman, bounces]
In-Reply-To: Your message of Fri, 13 Dec 91 11:31:58 -0500.
             <9112131631.AA03886@riesz> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Fri, 13 Dec 91 11:10:07 PST
Sender: jim%mycroft@rand.org
Status: OR


> jbaez@math.mit.edu writes:
> The D'Imperio book has NO COPYRIGHT on it so I will be unabashed about
> xeroxing it and mailing people copies.

Since Aegean is publishing it, I'd recommend having people order through
them.  If they're out, you should get permission anyway before cloning it.
Under the Berne Convention, you don't need to say Copyright (c) 1978 or
anything like that, if I understand it right.  In any case, let's not have
a copyright flame war in the list -- I think it should be a professional
issue rather than a legal one anyway.

> D'Imperio mentions that Tiltman's findings are the most worthwhile work to
> have been done on the Voynich.  Unfortunately they are unpublished.  She
> says she'll be glad to pass on stuff to serious scholars.  Any chance that
> someone (e.g.  Jim Gillogly :-) could track her down and get that stuff?

I wrote to her earlier this week to ask some of the questions we'd come up
with.  If she writes back I'll ask this one also.  I imagine some of the
work that's been done can't be put out to the masses because it uses
classified techniques.  Obviously Voynich decryptions wouldn't be
classified, but methods tried against it could give away indications about
the government's capabilities to potential intel targets.

> I've been getting "failed mail" bounces of my postings from
> raven.eklektix.com.  I presume my postings are getting through to the

I've been getting failed mail bounces on a few sites, especially in the
UK -- either of you UK types getting anything?  raven.eklektix.com hasn't
been bouncing for me yet.


Jim Gillogly

From gauss!rand.org!jim%mycroft Fri Dec 13 13:29:04 PST 1991
Received: by gauss; Fri Dec 13 17:05:24 EST 1991
Received: by inet.att.com; Fri Dec 13 17:05 EST 1991
Received: from mycroft.rand.org by rand.org; Fri, 13 Dec 91 13:29:07 -0800
Received: from localhost by mycroft.rand.org; Fri, 13 Dec 91 13:29:05 PST
Message-Id: <9112132129.AA12872@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: Split tall legged letters 
In-Reply-To: Your message of Sat, 14 Dec 91 08:14:06 -0500.
             <9112132114.AA15809@medici.trl.OZ.AU> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Fri, 13 Dec 91 13:29:04 PST
Sender: jim%mycroft@rand.org
Status: OR


> j.guy@trl.OZ.AU (Jacques Guy) writes:
> 
> I have a counter-proposition: transcribe the left half "q" and the
> right half "p" (lowercase letters). It's easy to remember: they do look
> like a tall "q" and a tall "p" standing on the line.

I like this idea.  I think P and F are both used this way.  Shall we
say [extended Q] -> p...q and [extended X] -> f...g?  Any opposed?

> I have found a serious problem with Currier's transcription: he uses
> both O (capital O) and 0 (zero), which can be quite confusing, depending
> on the character set installed on your machine. Ditto for his use of I 
> (capital I) and 1 (one). 

This doesn't bother me -- I'm planning to (eventually) do most of my human
looking at prints or Voynich-script displays, and the programs don't care.
I'd suggest using your new program to convert it each time before looking
at it.  I think the benefits of leaving Currier alone as much as possible
outweigh the benefits of tuning it in this case.

> I'd say: let's not rush into transcribing, but thrash out all the problems
> we encounter first. 

I have a big problem with this -- if we don't actually *try* transcribing
things, we won't understand the problems, and we'll have more problems to
thrash out anyway after we get there.

As long as we leave comments in to indicate where we had trouble, it should
be an acceptable amount of trouble to go back and fix it.

Jim Gillogly

From gauss!gauss.att.com!reeds Fri Dec 13 17:30:54 EST 1991
Received: by gauss; Fri Dec 13 18:13:55 EST 1991
Received: by inet.att.com; Fri Dec 13 18:13 EST 1991
Received: from research.att.com by rand.org; Fri, 13 Dec 91 14:36:32 -0800
Message-Id: <9112132236.AA27799@rand.org>
From: reeds@gauss.att.com
Date: Fri, 13 Dec 91 17:30:54 EST
To: voynich@rand.org
Status: OR

1. I like the extended Q -> p...q and X ->  f...g ideas, but what about
extended F and P (I saw one of these recently, I think...)?

2. I agree with Jim G. and disagree with Jacques about the need to
start transcribing soon.


From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Fri Dec 13 19:59 PST 1991
Received: by gauss; Fri Dec 13 23:26:29 EST 1991
Received: by inet.att.com; Fri Dec 13 23:26 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Fri, 13 Dec 91 20:03:02 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Fri, 13 Dec 91 20:07:44 PST
Date: Fri, 13 Dec 91 19:59 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: D'Imperio copyright.
To: voynich@rand.org
Message-Id: <CEF109E5D8DF006A88@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

Generally speaking, as I understand it, US Government publications are
not copyright, but in the public domain -- but then what the Berne
convention has done to this rule I don't know.  I will check --
But
    if it's a matter of  (a) reimbursing someone who has gone to the work
to put out an actual bound volume or (b) {will wonders ever cease?} of
the author actually getting some kind of ghreen (short if not long),
I'd certainly prefer to buy a text.

Books In Print as far as I know (but I only talked to the bookstore &
didn't double check myself) does not list it.  Should I try Forthcoming
Books?  Is there a release date?

Is this a copyright flame?  I hope not.  But I nevertheless see nothing
wrong with photocopying government publications if there's no
copyright and if neither (b) nor (a) holds.

From gauss!trl.OZ.AU!j.guy Sat Dec 14 07:49:17 EST 1991
Received: by gauss; Fri Dec 13 16:10:04 EST 1991
Received: by inet.att.com; Fri Dec 13 16:09 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Fri, 13 Dec 91 12:49:34 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA15791; Sat, 14 Dec 91 07:49:17 EST
Date: Sat, 14 Dec 91 07:49:17 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112132049.AA15791@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Currier's J vs 7
Status: OR


Look at folio f3v, which Jim has scanned into a gif file. On line 7, you have J ending the 
second word, and 7 near the end of the line. As I see it, the difference is that 7 is 
noticeably wider. The problem is: how wide is wide enough? I propose this. Look at the
left stroke, the one that looks like Voynich "I". If its foot is under the loop (right part),
call it J. If it ends left of the loop, call it 7:



    @@@   
   @@  @  <--loop
   @@  @
    @@@
@@ @@@@            foot under loop: J
 @@@    @
  @@     @
   @@     @
    ^     @
 foot     @
         @
        @




       @@@
      @@  @
      @@  @
       @@@
@@  @@@@           foot outside loop: 7
 @@@    @
  @@     @
   @@     @
          @
          @
         @
        @

From gauss!trl.OZ.AU!j.guy Sat Dec 14 07:59:47 EST 1991
Received: by gauss; Fri Dec 13 16:25:58 EST 1991
Received: by inet.att.com; Fri Dec 13 16:25 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Fri, 13 Dec 91 12:59:57 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA15796; Sat, 14 Dec 91 07:59:47 EST
Date: Sat, 14 Dec 91 07:59:47 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112132059.AA15796@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: a utility coming up, and some wild ramblings
Status: OR




(14 Dec 91, 2:17 yes, that's a.m.!)


First an announcement: I am working on a little program that will
translate Bennett's and Currier's systems into one another. Going from
Currier's system to Bennett's is trivial. The other way around not
quite so, and in fact, I want to write it so that it can translate from
just about any sensible system into another.  So give me a little
time.  Meanwhile, that means that if some of you can't stomach
Currier's, they can transliterate in Bennett's system (just add a "B"
or something to account for the letter that Currier transliterates
"J"), or have a go at meaking up their own (just don't go overboard).

And now, my turn to delve into wild speculations.

I was browsing through d'Imperio's book when I hit upon this table on
p.105, which I shall reproduce hereunder, both in Currier's and Bennett's
transliteration systems:

(Currier's Transliteration)

Roots                  Suffixes

OF-, OV-              -AD, -AN, -AM, -A3
OP-, OB-              -AR, -AT, -AU, -A0
4OF-, 4OV-            -AE, -AG, -AH, -A1
4OP-, 4OB-            -OR
S-                    -OE
Z-                    -C9, -CC9, -CCC9
8-                    -C89, -CC89, -CCC89
2-

Fig. 27 -- Tiltman's Division of Common Words into "Roots" and "Suffixes"
(Tiltman 1951)


(Bennett's Transliteration)

Roots                 Suffixes

OK-, OF-              -AU, -AN, -AM, -AIM
OH-, OP-              -AQ, -AIQ, -AIIQ, -AIIIQ
DOK-, DOF-            -AL, -AIL, AIIL, -AIIIL
CT-                   -OQ
ET-                   -CG, -CCG, -CCCG
S-                    -C89, -CC89, -CCC89
Z-


What does it all mean? This: take anything from the left column, whack on
anything from the right column, and, Abracadabra, Hocus Pocus, Presto
Shazam! here's a Voynich word. Now those of you who have studied Chinese (I
am sure there are some) will have recognized there something very similar
to the fanqie of the traditional analysis of Chinese words. Traditional
Chinese scholars analyze the Chinese syllable into initial, final, and
tone. The initial is the initial consonant of the syllable (sometimes
none), the final is the rest. In Mandarin (by which I mean the learned
variety of the Peking dialect, used for administrative purposes), there are
only about 400 possible different syllables, and a final can only end in a
vowel or either of two consonants: "n" or "ng". Further, "ng" never occurs
as an initial, and syllable-final "n" is acoustically quite distinct from
syllable-initial "n". What am I driving at? That the startling repetitive
patterns and co-occurrence restrictions of the Voynich language (IF it is
one), are compatible with an imperfect phonetic rendition of a language
such as Chinese. I do not remember whom I first bombarded with my pet
tongue-in-cheek theory about the Voynich manuscript. Was it Michael Barlow,
or Brian Winkel, editor of Cryptologia? Here it is: the Voynich manuscript
was written in Venice by two natives brought back by Marco Polo, probably
from China (tongue, stay firm in my cheek). They devised the alphabet after
what they had seen our writing, which they probably could not read. That
would not be an isolated incident: the Cherokee language is traditionally
written in a syllabary the symbols of which look strangely like Roman
letters. Its inventor, a Cherokee Indian by the name of Sequoia, devised it
last century. Sequoia was illiterate, but had figured out that there was a
correlation between the black squiggles in pale faces' books and the sounds
of their speech. So he set about devising a set of squiggles for his own
language, inspired from the pale faces' squiggles, ending up not with an
alphabet, but a syllabary. And a very good job he did too, entirely
adequate for recording the sounds of Cherokee. Chinese-like languages,
however, are another story. Their sound patterns (phonology, in our jargon)
are not easily amenable to sensible alphabetical or syllabic writing. And
how, a naive native speaker, would you represent tones in a writing system
to which you have just been exposed? That, together with tone sandhi (quite
extensive in those language families) would account for the strange
repetitive, but not quite exactly repeating, patterns of the Voynich
language. Let me give you one example. In Mandarin, the word for "Miss" is
made up of the same syllable twice repeated: jie3. The "3" here just means
that it is in the third tone: starting on a low pitch, going lower, then
rising sharply. However, when two third tones occur in succession, the
first becomes a second tone, rising sharply from middle pitch. Finally,
to confuse things further, the second "jie" becomes unstressed, loses its
tone, and is just uttered on a pitch slightly lower than the end pitch of
the first "jie"! Imagine now that you were a 13th century Chinese speaker
suddenly transported to England. The notion of an alphabetic writing system
would be novel to you. You would have great difficulties finding
correspondences between the sounds of English and those of Chinese. Sitting
down to write an encyclopaedia in alphabetic writing, you would soon be
confronted with delicate decisions: "'Map' is not too difficult to write,
it sounds about like the English word 'too'. So I write it 'too'. Uh, uh,
but 'earth' too, is 'too', and so is 'to spit', only they are in different
tones. Oh, well, not to worry, there can hardly be any confusion, given the
context, so I'll write them all the same. Whenever there is a serious risk
of confusion, I'll make something up. 'Too' (earth) is third tone, down
then up, so when the need arises, I'll write it 'too-oo' or something like
that, or perhaps I shall stick a tall squiggly letter in it".

Yes, I do have my tongue in my cheek when I say that the Voynich was
written by two Chinese speakers brought to Venice by Marco Polo. To be
honest, I estimate the likelihood of it to be so small as to be negligible.
At the same time, I think it infinitely more probable than Brumbaugh's and
especially Levitov's decipherments (John Baez, you're the mathematician, so
tell us: does it mean that I believe Brumbaugh's and Levitov's
decipherments to be worth exactly, precisely zilch and not even one
googolplexeth more?)



From gauss!trl.OZ.AU!j.guy Sat Dec 14 08:14:06 EST 1991
Received: by gauss; Fri Dec 13 16:53:25 EST 1991
Received: by inet.att.com; Fri Dec 13 16:53 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Fri, 13 Dec 91 13:14:16 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA15809; Sat, 14 Dec 91 08:14:06 EST
Date: Sat, 14 Dec 91 08:14:06 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112132114.AA15809@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Split tall legged letters
Status: OR


Jim Reeds suggests: "2. Sometimes the tall 2 legged letters like P and
F have their feet planted in different parts of a word, or even in
different words; the pedestal versions (Q and so on) have separate
``S'' shoes, one for each foot.  In this case (1) transcribe it as two
appearances of the letter (located where the feet land), but (2) put a
note in the comments section."

 I have a counter-proposition: transcribe the left half "q" and the
right half "p" (lowercase letters). It's easy to remember: they do look
like a tall "q" and a tall "p" standing on the line.

 I have found a serious problem with Currier's transcription: he uses
both O (capital O) and 0 (zero), which can be quite confusing, depending
on the character set installed on your machine. Ditto for his use of I 
(capital I) and 1 (one). 

 I'd say: let's not rush into transcribing, but thrash out all the problems
we encounter first. 

 Since I'll be posting soon a program to translate from any transcription
system into another, there would be no harm if we settled on our own(s):
we could just translate all of d'Imperio's transcriptions into it or the
other way around.

From gauss!cl.cam.ac.uk!Michael.Roe Sat Dec 14 16:58:29 +0000 1991
Received: by gauss; Sat Dec 14 12:07:47 EST 1991
Received: by inet.att.com; Sat Dec 14 12:07 EST 1991
Received: from swan.cl.cam.ac.uk by rand.org; Sat, 14 Dec 91 08:58:49 -0800
Received: from faeroes.cl.cam.ac.uk by swan.cl.cam.ac.uk 
          with SMTP (PP-5.7) to cl id <01655-0@swan.cl.cam.ac.uk>;
          Sat, 14 Dec 1991 16:58:34 +0000
Received: by uk.ac.cam.cl.lion (15.11/SMI-3.0DEV3) id AA15352;
          Sat, 14 Dec 91 16:58:30 gmt
Message-Id: <9112141658.AA15352@uk.ac.cam.cl.lion>
To: voynich@rand.org
Subject: Obtaining the D'Imperio book
Date: Sat, 14 Dec 91 16:58:29 +0000
From: Mike Roe <Michael.Roe@cl.cam.ac.uk>
Status: OR


As several posters have mentioned, the D'Imperio book is available from
Aegean Park Press. Many bookstores won't believe that it exists, as it's not
on their microfiche of books in print. If you know its ISBN, it's a lot
easier to convince a bookstore to order it:

D'Imperio, M. E. ``The Voynich Manuscript: An elegant enigma''
ISBN 0-89412-122-7 (library binding) ISBN 0-89412-038-7 (paperback?)

When I last checked, neither the British Library nor Cambridge University
Library had a copy. The Widener Library at Harvard had a copy (useful to
know if you're in Cambridge, Mass.)

Mike

From gauss!trl.OZ.AU!j.guy Sun Dec 15 08:31:46 EST 1991
Received: by gauss; Sun Dec 15 13:55:04 EST 1991
Received: by inet.att.com; Sun Dec 15 13:54 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sat, 14 Dec 91 19:49:27 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA16459; Sun, 15 Dec 91 08:31:46 EST
Date: Sun, 15 Dec 91 08:31:46 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112142131.AA16459@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Transcription problems: a solution perhaps.
Status: OR


(Sunday 15 Dec 91, 7:00)

Do I have a Xmas present for those of you with a PC and an EGA or VGA
display!

A Voynich font AND a transcription system such that you can write in
English and in Voynich with your favourite word-processor! All you have to
do is write the English in uppercase, and the Voynich in lowercase. The
Voynich letters are close enough to our Roman alphabet so that if you type
English in lowercase you'll still be able to read yourselves fairly easily.

At the core is a giftware font editor by Harald Thunem of Trondheim
(Norway), which I downloaded from an ftp site a few weeks ago. I never
thought I'd ever use it (I'll probably have to cough up some kroner, all
the more so that Harald Thunem distributed the Pascal source code with it,
which really is nice of him).

It all started last night, when I wrote this:

(Saturday 14 Dec 91, 19:25)
Time for a bit of Voynich. D'Imperio gives a frequency table of Voynich
characters p.106, which I reproduce here. The first column contains
Currier's alphabet. Since I gather from the e-mail that many do not have
access to D'Imperio's book where Currier's transcription is given p.97,
I have repeated the same letters in Bennett's transcription in the column
headed "B" (if you do not have Bennett's book, you will find his
transcription system in my article in the July issue of Cryptologia).


I have adapted D'Imperio's table:

1. I have replaced D'Imperio's absolute frequencies by percentages.
2. In the column headed "Krischer" D'Imperio mentions two occurrences of a
   character which is not a letter of Currier's system, but can be
   represented, in Currier's transcription, by II6. I have adjusted
   the count of I's and 6's, and the grand total accordingly.
3. I have replaced "Language A" and "Language B" by "Scribe A" and "Scribe
   B". Thanks to Currier, we know that there are two different
   handwritings, but that does not necessarily mean that they correspond to
   two different languages or dialects: the discrepancies between the
   letter frequencies in A an B could be due to different spelling habits.



           Currier    Currier   Krischer  D'Imperio
    B      Scribe A   Scribe B  f103-116  (Herbal,
           (Herbal)   (Herbal)            Astronomy)

4   D        2.48       2.30     4.76     2.03
O   O       19.21      12.29    14.88    18.69
8   S        7.55      11.19     8.29     7.35
9   G       10.51      13.69     9.47    10.44
2   Z        1.75       1.35     0.84     2.34
E   L        5.66       4.44     5.10     5.54
R   Q        4.53       4.43     4.10     5.35
S   CT      11.53       6.73     7.67     7.57
Z   ET       3.54       2.59     1.90     3.07
P   H        4.41       3.37     3.82     4.05
B   P        0.64       0.97     0.96     0.85
F   K        5.08       7.17     5.45     4.77
V   F        0.18       0.56     0.12     0.29
Q   CHT      1.41       0.46     0.27     1.47
W   CPT      0.36       0.11     0.14     0.25
X   CKT      0.73       0.90     0.31     0.58
Y   CFT      0.06       0.08     0.04     0.16
A   A        7.69       9.72    11.14     8.10
C   C        6.57      12.45    14.90     6.03
I   I        0.14       0.07     0.12
G   IL       0.03       0.01     0.00
H   IIL      0.01       0.00     0.00
1   IIIL     0.00       0.00     0.00
T   IQ       0.19       0.40     0.71
U   IIQ      0.07       0.21     0.22
0   IIIQ     0.03       0.02     0.02
D   U        0.32       0.03     0.08
N   N        0.70       0.65     0.78
M   M        3.89       2.56     3.12
3   IM       0.15       0.20     0.00
J   B        0.67       0.89     0.47
K   IB       0.05       0.04     0.02
L   IIB      0.01       0.01     0.02
5   IIIB     0.00       0.00     0.00
6   (B?)     0.11       0.06     0.06
7   (B?)     0.04       0.04     0.22

Letters
counted     11,709     11,168    4902   18,317



In Search of a Better Transliteration System

I have already remarked in an earlier posting on some shortcomings of
Currier's transcription system. From those frequencies, you see that
certain letters are extremely rare. By the time you encounter them when
transcribing from photocopies, you are likely to have forgotten their
Currier transliteration. There seems little point in particular in
memorizing letters G, H, and 1 since they occur so very infrequently and
can be built up from other elements. Ditto for T, U, and 0 and for K, L,
and 5. I now tend to see there a serious flaw in Currier's system, which
makes transcription harder and more susceptible to error. The use of 0 and
1, so easily confused with O and I, is also annoying.

In my opinion, a good transliteration system should be easy to memorize,
easy to remember, easy to type, and easy to read.

Easy to memorize:  there should be as few letters as possible to learn.
Easy to remember:  this is probably best achieved by choosing letters of
         the Roman alphabet that look as much as possible like the
         Voynich letters they symbolize.
Easy to type:  keep close to the home keys, minimize the use of the
         shift key.
Easy to read:  when you read a transcription, you should have little
         trouble visualizing what the Voynich original looks like, and
         no trouble translating back into Voynich letters.

After a bit of cogitation, I come up with an alphabet of just 18 symbols.
I avoided uppercase letters and shifted characters, however much tempted I
was to use the ampersand, which looks so much like a Voynich letter. I
broke up each of the four tall wiggly letters into two components: left and
right. This has the advantage of permitting to represent naturally those
few cases where such a letter "straddles" a word, with one foot here and
the other several letters away.

Now before you reach for your flame-throwers and turn me into a little pile
of ashes like you did Joan of Arc, I want you, please, to print out this
alphabet, and to write down the corresponding Voynich letters next to it.




Curr.   Benn.   Frogguy        Comments

4       D       4
O       O       o
8       S       8
9       G       9
2       Z       s         a mirror-image of the Voynich letter
E       L       x         close the top of the x for the Voynich letter
R       Q       2         z or ? could do, too
S       CT      ct
Z       ET      c't
P       H       qp        link the two for the Voynich letter
B       P       qg        g reminds me of the right half of that letter
F       K       jp        lp wouldn't be bad either
V       F       jg        lg ditto
Q       CHT     cqpt
W       CPT     cqgt
X       CKT     cjpt      or clpt perhaps?
Y       CFT     cjgt      or clgt perhaps?
A       A       a
C       C       c
I       I       i
G       IL      ix
H       IIL     iix
1       IIIL    iiix
T       IQ      i2
U       IIQ     ii2
0       IIIQ    iii2
D       U       v        the Voynich letter looks like v with a flourish
N       N       w
M       M       iw       (or m if you prefer)
3       IM      iiw      (or im if you prefer)
J       B       ig       it's i plus much the same g-loop as in jg and qg
K       IB      iig
L       IIB     iiig     yuck! but it hardly ever occurs anyway
5       IIIB    iiiig    ditto
6       (B?)    cg       looks like c plus the g-like loop
7       (B?)    i'g      a kludge, I know (*sigh*)


Finally, I want you to open Bennett at p.188 where there is a reproduction
of folio 79v (or look it up in Cryptologia) and to compare the first two
paragraphs with this transcription:

   qgox8c'tc89 oxjpo29 4oqpoxox oqpax89 9qpc8ox o2 oxo2ox
   4oqpcc89 4ojpctc9 4oqp9 xc'tc9 4ojpaw c'tc9 4o2ctc89
 8oxc'tcox ojpct9 4oqpaw c'tcc89 4ojpc'tc89 4ojpc89 oqpa2aig
 4oqpcc89 4oqpcc89 4oqpaw sox ctc89 2 ctc9 4ojp9 xctc89
 4oxjpcc89 4ojpc89 4oqpax saiw o29 4ojpc89 oqpcc89 xctcaig
 9ctccjptc9 2 aw ctc89 4ojpaw ctc89 ox c'tc89 8a2 9qpaig
 4ojpc'tc9 4ojpcc89 4oxjpc89 4ojpcc89 4ojpaw s oxctc89
 9ctc89 4oqpc9 ojpc89 qpc89ox c'tcc89 4ojpcc9 4oqpcc89 xox
 sa2 ox c'tcc9 4ojpcc89 4ojpcctc9 4ox
  qgctc89 xc'tccjptc89 4ojpcc9 4ojpaiw oxjp9 oqgctc89 qgctc89
  oxctcc9 xctc89 4oxjpcc89 4ojpaw ctcjpc9 oqpa2 oxjpaig


It's 22:20 now. I've spent three hours on that. I've earned a rest, don't
you think? But I'll proofread it first. Done. Now for a malt whisky. A
votre sante' a` tous!


Back to Sunday morning now. The fonts I designed had me change my
tansliteration system slightly. Here it is now:

Curr.   Benn.   Frogguy

4       D       4
O       O       o
8       S       8
9       G       9
2       Z       s
E       L       x
R       Q       2
S       CT      ct
Z       ET      c't
P       H       qp
B       P       qg
F       K       lp
V       F       lg
Q       CHT     cqpt
W       CPT     cqgt
X       CKT     clpt
Y       CFT     clgt
A       A       a
C       C       c
I       I       i
G       IL      ix
H       IIL     iix
1       IIIL    iiix
T       IQ      i2
U       IIQ     ii2
0       IIIQ    iii2
D       U       v
N       N       iv
M       M       iiv
3       IM      iiiv
J       B       iy
K       IB      iiy
L       IIB     iiiy
5       IIIB    iiiiy
6       (B?)    cy
7       (B?)    i-y


The first two paragraphs in folio 79v now become:

   qgox8c'tc89 oxlpo29 4oqpoxox oqpax89 9qpc8ox o2 oxo2ox
   4oqpcc89 4olpctc9 4oqp9 xc'tc9 4olpaiv c'tc9 4o2ctc89
 8oxc'tcox olpct9 4oqpaiv c'tcc89 4olpc'tc89 4olpc89 oqpa2aiy
 4oqpcc89 4oqpcc89 4oqpaiv sox ctc89 2 ctc9 4olp9 xctc89
 4oxlpcc89 4olpc89 4oqpax saiiv o29 4olpc89 oqpcc89 xctcaiy
 9ctcclptc9 2 aiv ctc89 4olpaiv ctc89 ox c'tc89 8a2 9qpaiy
 4olpc'tc9 4olpcc89 4oxlpc89 4olpcc89 4olpaiv s oxctc89
 9ctc89 4oqpc9 olpc89 qpc89ox c'tcc89 4olpcc9 4oqpcc89 xox
 sa2 ox c'tcc9 4olpcc89 4olpcctc9 4ox
  qgctc89 xc'tcclptc89 4olpcc9 4olpaiiv oxlp9 oqgctc89 qgctc89
  oxctcc9 xctc89 4oxlpcc89 4olpaiv ctclpc9 oqpa2 oxlpaiy

I sent Jim Gillogly a zip file FONTED.ZIP containing all Harald's package
plus a font file called VOYNICH.FNT.

To load the Voynich font just say LOAD VOYNICH. To unload it and go back to
your standard font, just give the DOS command MODE CO80 at the DOS prompt
(I'm assuming you have a color monitor; will it work too for monochrome?
Perhaps it's MODE BW80 or something like that). Try reading this message
with the Voynich font loaded and see!

From gauss!trl.OZ.AU!j.guy Sun Dec 15 09:42:27 EST 1991
Received: by gauss; Sun Dec 15 12:33:25 EST 1991
Received: by inet.att.com; Sun Dec 15 12:33 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sat, 14 Dec 91 19:48:43 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA16474; Sun, 15 Dec 91 09:42:27 EST
Date: Sun, 15 Dec 91 09:42:27 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112142242.AA16474@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Fonts/transcription latest
Status: OR


I've fiddled a bit with my Voynich font to make it more legible
when you write English in lowercase letters. That entailed two
changes in the transcription system:

Instead of g for the top right hovering squiggle of Currier's B and V
(Bennett's P and F), I'm using the double quote ("). Come to think of
it, perhaps I should have used a semi-colon instead, since it is of
such rare occurrence.


Instead of s for Currier's 2 (Bennett's Z) I used "z", since it is far
less frequent than s. Alternately, we could use ?, but it's a shifted
character.

I've redesigned the shapes of q, p, and l (the halves of the tall
wiggly characters with their feet on the line) so that they look much
more like English, and are still very much Voynich.

I've redesigned "a" so that it looks more Voynich-like, but still
recognizable English.

I'll post several alternative fonts to-morrow (Monday), or perhaps only
Tuesday (depends how it'll go).

From gauss!BINAH.CC.BRANDEIS.EDU!EVANS Sun Dec 15 21:52 EDT 1991
Received: by gauss; Sun Dec 15 22:06:58 EST 1991
Received: by inet.att.com; Sun Dec 15 22:06 EST 1991
Received: from binah.cc.brandeis.edu by rand.org; Sun, 15 Dec 91 18:53:59 -0800
Received: from BINAH.CC.BRANDEIS.EDU by BINAH.CC.BRANDEIS.EDU (PMDF #12080) id
 <01GE5O9LDI0W9S4R8F@BINAH.CC.BRANDEIS.EDU>; Sun, 15 Dec 1991 21:52 EDT
Date: Sun, 15 Dec 1991 21:52 EDT
From: Ronald Hale-Evans <EVANS@BINAH.CC.BRANDEIS.EDU>
Subject: Hale-Evans photocopies
To: voynich@rand.org
Message-Id: <01GE5O9LDI0W9S4R8F@BINAH.CC.BRANDEIS.EDU>
X-Vms-To: VOYNICH
X-Vms-Cc: EVANS
Status: OR

Argh! I've returned from visiting me sick old mum in hospital and now I see how
rumours start. Let me repeat: I do NOT have a full photocopy of the ms.. I have
a few pages that the Beinecke people gave me when I was a wee high school
student applying to Yale. I will be more than happy to make what I have
available to people on the list, but I am remodeling my library and my papers
are boxed away somewhere. I think I will take Mr. Baez up on his kind offer of
duplicating the pages I have and sending them out for free, because (1) we're
neighbours, and (2) *so* many people have requested pages, including one or two
requests from overseas.

Please be patient. I won't be able to get to the photocopies until after the
first of the year, and for all I know, they've already been duplicated in
readily-available research materials. I know, I know, I'd be eager to get my
hands on them too. Sorry about that. 

Ron Hale-Evans

From gauss!gmuvax2.gmu.edu!mmonk Mon Dec 16 00:28:32 0500 1991
Received: by gauss; Mon Dec 16 00:49:39 EST 1991
Received: by inet.att.com; Mon Dec 16 00:49 EST 1991
Received: from gmuvax2.gmu.edu by rand.org; Sun, 15 Dec 91 21:27:01 -0800
Received: by gmuvax2.gmu.edu (5.64/1.35)
	id AA15693; Mon, 16 Dec 91 00:28:32 -0500
Date: Mon, 16 Dec 91 00:28:32 -0500
From: mmonk@gmuvax2.gmu.edu (Martin Monk)
Message-Id: <9112160528.AA15693@gmuvax2.gmu.edu>
To: j.guy@trl.OZ.AU, voynich@rand.org
Subject: Re:  A better font
Status: OR



From gauss!trl.OZ.AU!j.guy Mon Dec 16 07:47:57 EST 1991
Received: by gauss; Sun Dec 15 16:07:29 EST 1991
Received: by inet.att.com; Sun Dec 15 16:07 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sun, 15 Dec 91 12:48:55 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA17012; Mon, 16 Dec 91 07:47:57 EST
Date: Mon, 16 Dec 91 07:47:57 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112152047.AA17012@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: A better font
Status: OR


(Sunday, 16:00 or thereabouts)

All right, folks, this is it. I sent a replacement for this morning's
Voynich font to Jim Gillogly. That replacement is the best font I could
come up with by far, so I won't send alternative fonts. I was especially
careful this time to design it so that it should be very easy, trivial
even, to read English written in those letters. This font was designed on a
VGA monitor, and it may not look so neat on an EGA. But I'll call it quits
for today. Here is the transcription system to go with it:

Currier   Bennett   Frogguy

    4       D       4
    O       O       o
    8       S       8
    9       G       9
    2       Z       z
    E       L       x
    R       Q       2
    S       CT      ct
    Z       ET      c't
    P       H       qp
    B       P       q;
    F       K       lp
    V       F       l;
    Q       CHT     cqpt
    W       CPT     cq;t
    X       CKT     clpt
    Y       CFT     cl;t
    A       A       a
    C       C       c
    I       I       i
    G       IL      ix
    H       IIL     iix
    1       IIIL    iiix
    T       IQ      i2
    U       IIQ     ii2
    0       IIIQ    iii2
    D       U       v
    N       N       iv
    M       M       iiv
    3       IM      iiiv
    J       B       ig
    K       IB      iig
    L       IIB     iiig
    5       IIIB    iiiig
    6       (B?)    cg
    7       (B?)    i-g

And here is the beginning of folio 79v again (just for fun, I really don't
feel like doing any hard work after all that):

   q;ox8c'tc89 oxlpo29 4oqpoxox oqpax89 9qpc8ox o2 oxo2ox
   4oqpcc89 4olpctc9 4oqp9 xc'tc9 4olpaiv c'tc9 4o2ctc89
 8oxc'tcox olpct9 4oqpaiv c'tcc89 4olpc'tc89 4olpc89 oqpa2aig
 4oqpcc89 4oqpcc89 4oqpaiv zox ctc89 2 ctc9 4olp9 xctc89
 4oxlpcc89 4olpc89 4oqpax zaiiv o29 4olpc89 oqpcc89 xctcaig
 9ctcclptc9 2 aiv ctc89 4olpaiv ctc89 ox c'tc89 8a2 9qpaig
 4olpc'tc9 4olpcc89 4oxlpc89 4olpcc89 4olpaiv z oxctc89
 9ctc89 4oqpc9 olpc89 qpc89ox c'tcc89 4olpcc9 4oqpcc89 xox
 za2 ox c'tcc9 4olpcc89 4olpcctc9 4ox
  q;ctc89 xc'tcclptc89 4olpcc9 4olpaiiv oxlp9 oq;ctc89 q;ctc89
  oxctcc9 xctc89 4oxlpcc89 4olpaiv ctclpc9 oqpa2 oxlpaig

q;c't89  ol;ct89  4olpc89 4oqpcc89 4olpc89 4oxqpc89 4oqpc89 olp9
8aiv a2 oxc'tc9 89qpaiv 4olpaiv ctccqpt9 olpcc89 4olpcc89 2o2
4olpcc9 4olpc89 olpcc9 4olpax c'tcc89 4olpcc89 2oxctc9 4olpcc89
9qpcc89 4olpcc89 4olpaiv oxlpcc9 ctolpaiv 89 qpcc9 4olpaiv
8ox c'tcc9 4ox oxlpc'tc9 4olpcc89 oxlpcc89 4ox olpaiiv ox9
4ox c'tcc9 ctox ox 8a2 4olpaiiv ctcclpc9 4olp9 oqpaiv o2aig
4olpaiv c'tcc9 4olpcc9 qpcc9 oqpcc9 xc'tcc9 4olpcc9 oxlpcc89 2ctc9
9lpaix c't9 4oxa2 c'tc9 4olpc89 4olpc89 4olpc89 8a2 oxlpaiv ctaig
8ctc89 xctc9 4cqp9 ct'c89 olpaiv 9lpccz oxlpc9 oqp9 c'tc9 4ox9
8 c'tc9 4olpax c'tcc89 c'tclp9 o2aiv oqpc't89 8aiv c'qptc2 a2o2
qpc'tc9 9lpcc9 2 ca pp aiv o2oiiiv
                {^^^^^ yup, strange, but that's what it looks like to me}

Makes it easy to proofread once you've loaded the font, doesn't it?


From gauss!trl.OZ.AU!j.guy Mon Dec 16 07:51:12 EST 1991
Received: by gauss; Sun Dec 15 16:22:34 EST 1991
Received: by inet.att.com; Sun Dec 15 16:22 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sun, 15 Dec 91 12:51:24 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA17016; Mon, 16 Dec 91 07:51:12 EST
Date: Mon, 16 Dec 91 07:51:12 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112152051.AA17016@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Folio 42V: transcription problems
Status: OR


I have been looking through the dozen or so photographic prints I had made
long ago from a microfilm of the VMS. Folio 42v provided me with food for
thought: seeing a couple of unusual letter sequences in it, I set about
transliterating it, expecting to learn something from the problems I was
sure to encounter. And indeed... Now, what follows is not a transcription I
made for later use, but an exploratory transcription, as it were, with the
only purpose of getting a feel for this thankless task. As I was
transcribing, I noticed that I could arrange the words to align like they
do in the manuscript, or almost, without too much effort. So I tried. I
even managed to sneak in a crude representation of part of the plant on
that page. Here it is. Consider it as a pure experiment, and pay particular
attention to my notes after the text proper. I think I have hit upon
something there. This transcription is meant to be looked at with the
font I announced in my posting entitled "A better font".


{Folio 42v  Single plant, tall thin straight stem, ending in }
{poppy-like pod surrounded by triangular stiff leaves, thistle-like}
{half-way up the stem five leaves, curvy, flowing upwards}
{the stem ends in four straight roots}

<042v01> q;cto cqp-o c'tcc9 4octo qpaiiv c'taxz  == ctox cro2 8aiv
<042v02> 8c'tc9 qpctc9 9 lpctc9 ctqpct9 cq;t9 8av== 8aiv oqpox 8aiiv
<042v03> c'to ctoqpo89 ctoqpox olp9 ctox ctox    == 8xctcqpt9 oqpo9
<042v04> 4oqpct9 ctlpct9 za289   ====== o8aiix   == 9lpct9 olpox8cg
<042v05> ctolpcco 2 ctc9 ctcqpav=========za29    == olpox ctcqpaiiv
<042v06> olpcolpca2 ctcoqpcta2 =========== z9    == zaiiv cqpta28aig
<042v07> ctolp c'tcalpc9 lpccc98==========       ==  == ctclpcczc't9
<042v08> zo8ax zaiiv ===        ===========      == ====     ======
                     ======    =============   =========  =========
                       ==========================================    ====
                         ======================================   ======
                            ==========================================
                               ================  ===================
                                   ===========   ================
                                     ==========  =========
                                         =============
<042v09> q;pc'tco2 c'to  cta2 ctclpctc9  4olpo8  == c'to cq;tco zct9
<042v10> zcta2 olpctc9  ct9 c'tox oxo2 ctcaiiv   == cqpct9 qpctczz8
<042v11> 4o aiiv 4olpcc9  lpctox ctcax oqpcox8   == c'tc9 qpcaiivig
<042v12> ctoqpctcc9  ctox  ox ctaiiv oqpccaiiv   == ctlpct9 qpaiiv9
<042v13> c'to  ctc9  qpco2 ctc9  ctccqptt9 8aiiv == ctolpcc9 8aiiv
<042v14> 4oqpcc9  ct9 lpct9 cqpt9 cqpt'ccz ctox9 ==  za2a2  a2ax
<042v15>  ctolpc9 ctox  olpc9                    ==  olpo2 o8cc9


{Line 1. The ; of the initial q; is an embellishment, looking a bit like}
{    a hydrogen balloon in the shape of a fat woman, tied to the q stem}
{    "cqp-o": one would expect the more common form cqpt, but the loop}
{    of the "t" is closed, looking definitely like "o", clearly linked }
{    to the preceding "c" by a thick horizontal line, represented here}
{    by "-"}
{Line 2. "cq;t9 8av": the ";" is not the regular form, but looks like}
{    the digit "3", giving a compound very much like "q3"}
{Line 14. "cqpt'ccz": the "c" preceding "qp" is definitely connected}
{    to the letter immediately following "qp", and that letter looks}
{    like "c'" or "z". But since it is connected to a preceding "c"}
{    the transliteration alphabet leaves us only with "t", hence }
{    the present transcription: "cqpt'ccz". I find this unsatisfactory.}
{    It could well be that a "c" connecting to a following letter, }
{    optionally through intervening tall wiggly letters, does actually}
{    constitute a separate letter. In that case, what I have transcribed}
{    "cqp-o" in Line 1 would be that letter, followed by "qpo", and this}
{    strange sequence "cqpt'ccz" would in fact again be that letter,}
{    followed by "qpzccz". I think we should seriously consider that
{    possibility.}






From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Mon Dec 16 08:45 PST 1991
Received: by gauss; Mon Dec 16 13:12:12 EST 1991
Received: by inet.att.com; Mon Dec 16 13:10 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Mon, 16 Dec 91 08:57:54 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Mon, 16 Dec 91 09:01:48 PST
Date: Mon, 16 Dec 91 08:45 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Obtaining the D'Imperio Book
To: voynich@rand.org
Message-Id: <CCF3BD70B0FF4022A1@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

 I have just checked with our GovDocs librarians, who insist that to their
knowledge the Berne Convention copyright regs do not affect the public domain
status of things published by the US Gov't or its agencies.

At the same time (or almost), I see Mike Roe's message about the APP
edition of the D'Imperio still being available.  So -- why not?  When 
the book store opens at 9, I'll give them a call and see if they'll accept
an order.

We shal see...

From gauss!SEI.CMU.EDU!firth Mon Dec 16 15:48:10 0500 1991
Received: by gauss; Mon Dec 16 18:10:13 EST 1991
Received: by inet.att.com; Mon Dec 16 18:10 EST 1991
Received: from BP.SEI.CMU.EDU by rand.org; Mon, 16 Dec 91 13:58:45 -0800
Received: by bp.sei.cmu.edu (5.64/2.5)
        id AA27012; Mon, 16 Dec 91 15:48:10 -0500
Date: Mon, 16 Dec 91 15:48:10 -0500
From: firth@SEI.CMU.EDU
Message-Id: <9112162048.AA27012@bp.sei.cmu.edu>
To: voynich@rand.org
Subject: Personal Notes (1)
Status: OR

Notes on the Voynich Manuscript - Part 1
----------------------------------------

This is my first attempt to contribute to the Voynich mail group, 
so please be patient if I repeat old stuff.  To date, my information
comes only from

(a) an initial reference in Colin Wilson's silly book 'The Philosophers
Stone'

(b) a brief discussion in 'Labyrinths of Reason', already cited

(c) Brumbaugh's collection of articles 'The Most Mysterious Manuscript'

(d) this group

(e) the partial ASCII transcription available to us.

Not much to go on.  In particular, this means I've actually seen only
12 pages of the document in photocopy, and that unfortunately often
reduced, badly copied, and monochrome.

First Impressions
-----------------

What seems obvious from all this?  First, the document seems to be
a compendium of herbal, astrological, and maybe medical data, if
one judges by the illustrations.  It seems to me that the illustrations
were written before the text, and the text often labels figures, so
is probably about them.  Secondly, the writing is a clear, cursive
script that flows from left to right and top to bottom, as is evident
from the margins.

The text I have seen is divided into short paragraphs, each introduced
by a larger and slightly more ornate letter - a common mediaeval style.
On two of the photocopies, the reproduction is good enough to show the
density of the ink, and this, plus the cursive appearance, shows that
the text was probably written with reasonable speed - the speed of a
fair copy, not that of notes, nor that of a laboriously transcribed
cypher text.  (See, for instance, fol 85r on p 54 of Brumbaugh.)
Incidentally, it is asserted, and my small sample confirms, that there
are no corrections in the document.

The document is European.  This is suggested by its provenance, and
confirmed by several features of style.  The chubby ladies in its
margins are Caucasian; the zodiac it uses is Western; the stars and
constellations identified are those of the Northern Hemisphere; the
plants are European except for two disputed ones.

Key Issue
---------

As I see it, the key issue in attempting to understand this document
is to classify it.  The limited background material I've read suggests
three possibilities

G: gibberish   - the thing is a forgery without meaning
C: cypher text - the thing is an intricate cypher
P: plain text  - the thing is a real book intended to be read

There is also the possibility of G+C: the document is mostly null
padding with a small amount of encyphered information, perhaps one
letter per word or one word in ten.

After a little computer analysis, which I'll report on later, I've
decided it is very unlikely to be cypher text, and that any deliberate
concealment is largely incidental to its real purpose.  What I'd like
to put forward here is some reasons why it might be plain text.

Why Can't we Read it?
---------------------

If the Voynich manuscript is plain text, why can't we read it?  One
obvious answer is: for the same reason we couldn't read the Pyramid
Texts or the Dresden Codex - because the language, or script, or
both, had by historical accident passed out of use.  After all, only
three Mayan codices survived, so maybe a library of Voynich stuff
was burned before the fall of Mont Sejur.  In other words, the book
was written to be read, in a script and language well known to its
original reader community, whoever they were, and by purest accident
no other such book has survived.

I am not convinced of this, by any means, but at least think we
shouldn't dismiss the possibility.  Now, it is also possible the
script and language were deliberately invented, but even so the
prime motive need not have been concealment.  Consider the script.
It has clearly been devised with some care for its legibility and
ease of use, the manner in which the symbols join together confirms
that.  The motive could have been to create a cursive "shorthand";
more plausibly, it may simply have been to create a script easier
to read than latin minuscule, which the Voynich script certianly is.
And that's historically plausible too, for the italic cursive hand
was devised for just that reason.

Even a synthetic language need not have been for secrecy.  After
all, nobody would entertain for thirty seconds the idea that
Esperanto was devised to be the secret speech of an International
Cathar Conspiracy, and it troubles me that we accept such ideas
about the Voynich manuscript so uncritically.  What is its subject
matter?  Plants, stars, chemicals and the like.  It is entirely
reasonable that somebody should want to devise an international
synthetic language to discuss such things - after all, in this
twentieth century, we do discuss these subjects in an international
terminology that uses what are to all practical purposes synthetic
languages, such as the Geneva names for organic compounds, or the
Linnaean taxonomy for living organisms.  Something like "4oxlpcc89
4olpc89" could merely be an earlier philosopher's attempt at
"Echinus europaeus" or "C6H5CH2COOH".

Now, given the temper of the times, it is probable that the secrecy
of a specialist script or language was a welcome side effect, and
that it was used to mystify the uninitiated.  However, I find it
also plausible that this was a side effect, and that the main purpose
of the Voynich document was to communicate to those who could read it.
The probable existence of two scribes - indeed, I believe of two
scribes and at least one other author - lends support to this.

Finally, look at the document.  I could not write english cursive
this well, this consistently.  And I certainly couldn't copy a
text this long without error, unless I clearly understood its
meaning.  So, maybe the thing is a forgery, and there are no
corrections because there is no meaning to correct.  But even to
generate this forgery, somebody had to spend an immense amount of
time practicing a nonsense script, to become this fluent.  And I
find that very hard to believe; there are far easier ways to con
a mark out of a couple of hundred ducats.

Gold Bugs are Red Herrings
--------------------------

Finally, one trivial point. With regard to the systematic analysis
of the text, a document written in an unknown alphabet is
indistinguishable from one written using a simple substitution
cypher, and I simply lump both possibilities together as my
hypothesis P.  For example, if Capt Kidd's note had been written
using the hebrew alphabet, it could have been broken by the exact
same techniques as Poe actually describes in the story.  It is
irrelvant how the script was devised or to how many people it
was known; what matters is that we don't know it and must somehow
reconstruct its rules.  As to what those rules are - what are the
morphemes and phonemes of the script - that's a hard question.

Robert Firth

From gauss!trl.OZ.AU!j.guy Mon Dec 16 16:15:06 EST 1991
Received: by gauss; Mon Dec 16 00:54:36 EST 1991
Received: by inet.att.com; Mon Dec 16 00:54 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sun, 15 Dec 91 21:16:22 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA17946; Mon, 16 Dec 91 16:15:06 EST
Date: Mon, 16 Dec 91 16:15:06 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112160515.AA17946@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: some thoughts
Status: OR


On of the strange properties of the Voynich language/script is the constant
recurrence of the same word-initials.

I let myself imagine that the writing system was Indian (as in "Sanskrit")
but used letters inspired from the Roman alphabet. Indian writing systems
have, for us, very strange quirks:

1. The vowel "a" is not written.
2. Certain vowels are written BEFORE the consonant, some AFTER, some
   half before, half after, some ABOVE it, some UNDER it.
3. Since "a" is not written, a special sign is used to signal that a
   consonant does not have a vowel going with it.

To give you an example, it is as if "ROTARY" were spelt "ERATYR":

E-A  being O so: ERA = RO 
Y-   being Y so: YR  = RY
T    being TA since "A" is not written.

"NOT" would be spelt something like "ENAT*", with the asterisk indicating
that the T is alone, without an "A".

Many languages are written in such a system: all the languages of India
and Ceylon, Thai, Mon, Laotian, Burmese, Cambodian, Javanese, Balinese,
Tibetan, and probably more that I never even heard of.

Then I wondered: perhaps the tall, wriggly letters that often intrude
in Currier's S and Z (Bennett's CT and ET) are consonants and S and Z
(CT and ET) are vowels, one half of which is written to the left of the
consonant, the other to its right. And perhaps the same letters (S and Z),
when not separated by an intevening letter, are different letters 
altogether! Perhaps the group "4O" which occurs all the time at the
beginning of words is actually a vowel written BEFORE ITS CONSONANT.

Perhaps the "9" which ends so many words just marks the absence of a
vowel.

So, keep your minds WIDE open. It could be tough.


From gauss!SEI.CMU.EDU!firth Mon Dec 16 17:04:48 0500 1991
Received: by gauss; Mon Dec 16 18:24:37 EST 1991
Received: by inet.att.com; Mon Dec 16 18:24 EST 1991
Received: from BP.SEI.CMU.EDU by rand.org; Mon, 16 Dec 91 14:05:04 -0800
Received: by bp.sei.cmu.edu (5.64/2.5)
        id AA27133; Mon, 16 Dec 91 17:04:48 -0500
Date: Mon, 16 Dec 91 17:04:48 -0500
From: firth@SEI.CMU.EDU
Message-Id: <9112162204.AA27133@bp.sei.cmu.edu>
To: voynich@rand.org
Subject: Notes (2)
Status: OR

Notes on the Voynich Manuscript - Part 2
----------------------------------------

This is the report of a very brief initial computer analysis of
the ASCII text.  While I intend to go over to the Frogguy rules
of transcription as soon as possible, this analysis was done in
a couple of lunch hours some time ago, so uses the Currier rules.

Separate A and B
----------------

First, I separated the A and B sections, creating simple text
files VA and VB.  The reason for this is pretty obvious: if the
two sections are in different languages or use different cyphers,
I've made my task far harder by conflating them; but if they
are in fact the same, I've not made my task much harder by
dividing them, since any method that can crack a 100-page document
can surely crack a 50-page half document.

Letter Counts
-------------

Secondly, the obvious first step is to count the letter frequencies.
Here is the list for the VA and VB texts, including spaces (sp) and
new lines (nl):

VA
--

sp:  5861
O:   5152
9:   2878
S:   2852
A:   2050
8:   2011
E:   1523
C:   1500
R:   1320
F:   1296
P:   1166
nl:  1263
M:   1009
Z:    903
4:    631
2:    482
Q:    432
B:    200
J:    198
N:    193
X:    163
W:     89
D:     76
V:     60
T:     56
I:     47
3:     34
6:     34
U:     22
Y:     20
K:     11
7:      8
0:      5
G:      4
H:      4
L:      4
,:      3

VB
--
sp:  7977
C:   5688
9:   5308
O:   5283
8:   3990
A:   3065
E:   2839
F:   2794
S:   2206
4:   1933
Z:   1367
R:   1366
P:   1228
nl:  1165
M:    708
N:    556
2:    523
B:    303
X:    300
Q:    171
J:    152
V:     96
T:     78
3:     48
U:     34
W:     33
D:     22
6:     19
Y:     13
G:      9
I:      8
7:      5
K:      5
H:      3
L:      3
0:      2
5:      2

And, for comparison, here are the most common letters in a sample
of english text (E) and latin text (L).

E
-

sp:  5182
e:   2730
o:   1717
a:   1683
t:   1681
n:   1504
r:   1469
s:   1409
h:   1401
i:   1315
l:   1010
d:    901
f:    743
u:    558
nl:   548
c:    519
w:    469
m:    449
b:    413
y:    344
g:    333
p:    260
v:    221
k:     84
j:     45
x:     36
q:     17
z:      8

L
_

sp:   132
e:    105
i:     89
t:     89
u:     71
m:     62
a:     57
r:     56
s:     56
o:     55
n:     54
c:     41
nl:    39
p:     28
d:     24
l:     23
f:     12
v:     12
x:      7
b:      6
g:      6
q:      6
h:      5
j:      2

Well, the patterns are similar.  The Voynich A text seems to have about
20 to 25 common symbols, with frequencies that follow the Zipf law.  The
average word length is 5 symbols, but I now find the Currier transcription
uses single letters for what seem in the original to be compound symbols.
If those compounds are written in full, the word length is nearer 7 than 5.

Note also that VA and VB differ considerably.  Look at the frequency of 'O'
in VA and VB, and compare the frequency of 'o' in E and L.  However, this
doesn't prove they are different languages or use different cypher keys.
At that time, there were many dialects of our modern languages, and few
consistent schemes of spelling.  To take an example almost contemporary
with the VM, consider 'The Romaunt of the Rose'.  This is all in English,
but the two major parts are by different authors, use very different
dialects, and employ different spelling rules.  (Incidentally, nobody
seems to have pointed out that one of those authors, Geoffrey Chaucer,
was pretty interested in alchemy, medicine, and astronomy...)

Word Counts
-----------

The next step would be word counts.  However, are we sure the text is
divided into words?  My answer is yes, I am sure.  If the spaces were
inserted at random, or by some rule intended to conceal cypher groups,
then the initial and final letter frequencies should be close to the
medial frequency.  This is definitely not the case; there is a clear
pattern of preferred initial and final letters, just as in natural
language.  This being so, there are some significant data in these
counts.  Consider this word, from VB, where I put the count in brackets:

	SC89 [250]

This is a pretty common word, like, perhaps "SOME" in english.  Now,
many cypher schemes might encode "SOME" as "SC89".  But anything more
complex than a Gold Bug cypher would not also encode "winsome",
"frolicsome", "meddlesome" &c so they also ended in "-SC89".  Now
look at

	40ESC89 [12]
	8SC89 [15]
	BSC89 [18]
	ESC89 [57]
	0BSC89 [17]
	OESC89 [30]

That's another 150 occurrences of the same symbol group, which I
consider a pretty clear refutation of the notion that this is some
immensely intricate cypher.  This stuff feels like language.  Now,
do we have here the equivalent of "domina femina carmina candida",
or rather the equivalent of "eeny meeny miney mo"?  For nonsense,
too, shows the morphemic regularities of language.  I don't know;
but for the present I'm pretty sure hypothesis C - cypher text -
is a very distant third horse in this race.

Word Analysis
-------------

Well, let's run for a while with hypothesis P - we have here plain
text that happens to be written in an unknown script, transcribing
an unknown language.  What do we do?  We do what Michael Ventris
did when faced with this exact problem in the form of the Cretan
Linear B script: look for regular patterns suggestive of first
phonemes, and then grammar.

How is this language encoded?  Over 95% of the text is encoded using
20-odd symbols.  Now, it is just possible that the underlying language
has three vowels and seven consonants, for a 24-symbol syllabery, but
I don't believe it.  This is an alphabet, and each symbol is a consonant,
vowel, or possible consonant cluster or diphthong.

Does it show grammatical regularity?  You bet.  We've already seen
that "SC89" is a common word termination; there are also:

	FAM	[13]	FAN	[16]	FAR	[17]
	OEFAM	[12]	OEFAN	[24]
	OFAM	[56]	OFAN	[50]	OFAR	[44]
	OPAN	[18]	OPAM	[27]	OPAR	[38]
	ORAM	[10]	ORAN	[10]
	4OFAM	[96]	4OFAN	[154]	4OFAR	[66]
	4OPAM	[15]	4OPAN	[22]	4OPAR	[17]

from the same VB text.  Look, please, at the pattern of those word
counts.  Those frequencies can be explained as a combination of root
frequency ("4OF-" is much more common than "OP-") and ending frequency
("-AM" is slightly less common than "-AN").  This is precisely what
one would expect of a technical document written in a language with
end inflections.

But there are some peculiarities.  One is that the inflection scheme
is very, very consistent, far more so than latin.  There seems to be
only one major declension or conjugation.  Another is that what seem
to be grammatical endings are also words in their own right:

	AM	[70]	AN	[24]	AR	[58]

as if latin had words "us", "um", "os" to match its endings.  That
may have been the case in proto-indo-european, but it's not the
case with greek, latin or most modern european languages.  It would
also be surprising in a synthetic language devised by a european;
such languages tend to follow the same patterns as the natural
languages known to the devisor: as examples I cite Volapu"k and
Esperanto.

A third point, and an unpleasant one, is that the VA and VB texts
seem to have similar beginning bits (roots?) but different ending
bits (inflections?), and that shouldn't happen either.  If two
dialects diverge that much in respect of grammar, there should be
at least a vowel shift in the roots as well - compare german and
dutch to see an example.

Finally, if those bits are roots, they are very short.  You would
expect things like "ferr-", "cupr-", "chalc-", "aur-", "argent-',
and such, not roots of one or two letters.  Maybe they are encoded
in some compact notation, as if one would write

	Misce feum & cuum in Xbile

abbreviating "ferrum", "cuprum" and "crucibile" in the obvious way.
One could do the same with plant and star names, of course.

At which point, I ran out of time and ideas.  In the past couple
of weeks, I've had more ideas, but no time to test them.  If you
find any of this useful, I'd be happy to hear your ideas.

Robert Firth


From gauss!math.mit.edu!jbaez Mon Dec 16 17:32:35 EST 1991
Received: by gauss; Mon Dec 16 17:53:55 EST 1991
Received: by inet.att.com; Mon Dec 16 17:53 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Mon, 16 Dec 91 14:36:09 -0800
Received: from banach (BANACH.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA01162; Mon, 16 Dec 91 17:31:37 EST
Received: by banach; Mon, 16 Dec 91 17:32:35 EST
Date: Mon, 16 Dec 91 17:32:35 EST
From: jbaez@math.mit.edu
Message-Id: <9112162232.AA00444@banach>
To: voynich@rand.org
Status: OR

Since it appears that the D'Imperio *is* easily available, and probably
copyrighted, I rescind my offer to send it out free to anyone who asks.
Jim G., recall, says it's being published by Aegean Park Press.  They're
at PO Box 2837, Laguna Hills, CA 92654-0837, 714-586-8811.   I figure
that as this press has been kind enough to publish so many texts on
the Voynich they deserve to earn a little money for their contribution
to the cause.

I will wait for Ron Hale-Evans to get his pages and announce when I get
them; *then* ask for copies if you want them.

I attempted to transcribe folio 93r of the Voynich this weekend, and ran
into some questions.  Some were nicely anticipated by Jim R's post.
For those who have access to Brumbaugh's book, 93r appears there on
page 80.  It is the page with a picture of a "sunflower" that was used as
evidence for a post-Columbian origin of the Voynich.

1)  I was unable to distinguish between a 7 versus J.  Maybe we should
scratch one of those designations?  (For Jacques Guy's contribution
to the "7/J" issue -- just received -- see my grumpy comments below.) 
I note that D'Imperio's transcription contains both 7's and J's.
Has anyone seen a description of Currier's system *other* than that in the
back of D'Imperio?  (In that reference 7 and J look identical, but perhaps
she simply miscopied?)  On folio 93r I find examples of a character that
resembles a 7/J but is notably different.  For example, this character
is the last one on the 4th line.  Perhaps this is the "6" in Currier's
system?  Also take a look at the last character on the 4th from last
line.  This could be "6" or a "7/J".   

2)  I have trouble with R's versus S's.  There are what seem to be clear
examples of one or the other form, but then there are ambiguous cases.
Both R's and S's are examples of "enders," i.e. they tend to end words.
Having observed this, one tends to want to include them at the end of a
word even when -- as happened frequently on 93r -- they appear with a
slight space in front of them, which could lead one to take them as a
separate "word".  Spacing is in general a nuisance.

3)  I note a tendency to be less willing to note questionable
characters, and more eager to just take a stab in the dark, as one tires
of transcribing.

I'll post my transcription or send it to Jim G when I type it in.

Note that while it's good to try to reach conformity on these
transcription issues (issues of classification), so that the
transcription we produce will be of one piece, there will NECESSARILY be
lots of ambiguous cases one cannot resolve.  Normally these ambiguities
are resolved by knowing what the word might MEAN.  I do not want to get
into a baroque transcription scheme in a hopeless attempt to resolve all
ambiguities.  For example, Jacques' proposed distinction between 7 and J
seems so subtle that by this standard there are several hundred distinct
characters.  


From gauss!rand.org!jim%mycroft Mon Dec 16 14:48:54 PST 1991
Received: by gauss; Mon Dec 16 18:11:49 EST 1991
Received: by inet.att.com; Mon Dec 16 18:11 EST 1991
Received: from mycroft.rand.org by rand.org; Mon, 16 Dec 91 14:48:58 -0800
Received: from localhost by mycroft.rand.org; Mon, 16 Dec 91 14:48:55 PST
Message-Id: <9112162248.AA18818@mycroft.rand.org>
To: voynich@rand.org
In-Reply-To: Your message of Mon, 16 Dec 91 17:32:35 -0500.
             <9112162232.AA00444@banach> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Mon, 16 Dec 91 14:48:54 PST
Sender: jim%mycroft@rand.org
Status: OR


> jbaez@math.mit.edu writes:
> that as this press [Aegean Park] has been kind enough to publish so many
> texts on the Voynich they deserve to earn a little money for their
> contribution to the cause.

...  And maybe if we buy enough of D'Imperio (and Levitov?) from them
they'll be more willing to publish the complete picture book if we can get
clearance from Beinecke.

> 2)  I have trouble with R's versus S's.  There are what seem to be clear
> examples of one or the other form, but then there are ambiguous cases.
> Both R's and S's are examples of "enders," i.e. they tend to end words.
> Having observed this, one tends to want to include them at the end of a
> word even when -- as happened frequently on 93r -- they appear with a
> slight space in front of them, which could lead one to take them as a
> separate "word".  Spacing is in general a nuisance.

(I assume you mean R versus 2.) I've been coding the "slight space" as a
comma, even when I'm convinced I "know" that the following stuff is part
of the same word from staring at a lot of similar samples.

> 3)  I note a tendency to be less willing to note questionable
> characters, and more eager to just take a stab in the dark, as one tires
> of transcribing.

Time to go out for a beer!  Transcription is indeed tedious, but decrypting
noisy data is much much harder than decrypting clean data, and knowing where
the noise is is very important.

> I'll post my transcription or send it to Jim G when I type it in.

If you send it to me, I'll add it to the file "voynich.now" for everybody
to pick up from rand.org:pub/jim .  That'll be less work on the mailer.
There are about three dozen people on the list now, and I suspect some
are here just to see what progress is being made rather than to participate
in the grungy details.  (Everybody should feel welcome to participate in
the discussions -- this isn't meant as a put-down.)

> Note that while it's good to try to reach conformity on these
> transcription issues (issues of classification), so that the
> transcription we produce will be of one piece, there will NECESSARILY be
> lots of ambiguous cases one cannot resolve.  Normally these ambiguities

Amen!

	Jim Gillogly

From gauss!math.mit.edu!jbaez Mon Dec 16 18:20:26 EST 1991
Received: by gauss; Mon Dec 16 18:52:45 EST 1991
Received: by inet.att.com; Mon Dec 16 18:52 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Mon, 16 Dec 91 15:23:55 -0800
Received: from banach (BANACH.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA01452; Mon, 16 Dec 91 18:19:26 EST
Received: by banach; Mon, 16 Dec 91 18:20:26 EST
Date: Mon, 16 Dec 91 18:20:26 EST
From: jbaez@math.mit.edu
Message-Id: <9112162320.AA00584@banach>
To: voynich@rand.org
Subject: R's vs 2's
Status: OR

In my last post I said I had trouble with R's vs S's.  I meant
R's vs 2's.  I will not post my transcription but rather send it
to Jim G., which should become the standard modus operandi as work
proceeds.  

Jim R. says he bought the last copy of D'Imperio from Aegean Park
Press 2 weeks ago.  I am presently undecided as to how this should 
affect my recently made, recently rescinded offer to send copies to
people.  I emphasize, however, that this is a MUST READ for any
serious voynichologist.

jb

From gauss!rand.org!jim%mycroft Mon Dec 16 16:14:51 PST 1991
Received: by gauss; Mon Dec 16 19:53:22 EST 1991
Received: by inet.att.com; Mon Dec 16 19:53 EST 1991
Received: from mycroft.rand.org by rand.org; Mon, 16 Dec 91 16:14:55 -0800
Received: from localhost by mycroft.rand.org; Mon, 16 Dec 91 16:14:52 PST
Message-Id: <9112170014.AA19118@mycroft.rand.org>
To: voynich@rand.org
Subject: Good news from Aegean Park Press
Date: Mon, 16 Dec 91 16:14:51 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

I just talked to Wayne Barker at Aegean, and he said that although they
are indeed sold out of the pre-printed copies of D'Imperio, as they get
orders for it they will xerox the one they've got left, velo-bind it, and
send it out.

He also said that we can get 20% off the list price, which he forgot but
declared to be $24.80.  Add $2 shipping and handling.  Besides the address
and phone that John posted, they've got an 800 number:  1-800-736-3587.
When you call or write, say that you're on the Voynich Mailing List,
mention my name, and say that you'd like the 20% discount.

The 20% will work for anything in their inventory, including the Levitov
book on his proposed solution (but read Jacques' review in the ftp archives
if you're thinking of ordering).

More good news: I asked him whether they'd be willing to publish the Voynich
if we can get prints or microfilm, and he said absolutely -- he was quite
interested in doing it.  Microfilm would be fine: he can get that printed
and offset and the whole bit.  So all we need now is a good quality film
and we're in business.

	Jim Gillogly

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Mon Dec 16 16:15 PST 1991
Received: by gauss; Mon Dec 16 19:43:36 EST 1991
Received: by inet.att.com; Mon Dec 16 19:43 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Mon, 16 Dec 91 16:18:14 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Mon, 16 Dec 91 16:22:57 PST
Date: Mon, 16 Dec 91 16:15 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Re: R's vs 2's
To: jbaez@math.mit.edu, voynich@rand.org
Message-Id: <CCB4E912BADF404480@MAX.U.WASHINGTON.EDU>
X-Envelope-To: jbaez@math.mit.edu, voynich@rand.org
X-Vms-To: IN%"jbaez@math.mit.edu"
X-Vms-Cc: IN%"voynich@rand.org"
Status: OR

Well, I managed to convince the university bookstore to order one
for me despite its absence from BIP; I think I will write to APP and
see what else they have available.

If worse comes to worse, I will go argue with interlibrary loan...

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Mon Dec 16 21:27 PST 1991
Received: by gauss; Tue Dec 17 00:50:51 EST 1991
Received: by inet.att.com; Tue Dec 17 00:50 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Mon, 16 Dec 91 21:29:37 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Mon, 16 Dec 91 21:34:20 PST
Date: Mon, 16 Dec 91 21:27 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Good News from APP
To: voynich@rand.org
Message-Id: <CC8944164E5F404530@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

An 800 number yet!  They are going to get a call from me tomorrow 
morning.  Thanks for clearing all this up.
--RJB

From gauss!castrov.cuc.ab.ca!wuth Mon Dec 16 23:31:30 MST 1991
Received: by gauss; Tue Dec 17 05:44:43 EST 1991
Received: by inet.att.com; Tue Dec 17 05:44 EST 1991
Received: from fsa.cpsc.ucalgary.ca by rand.org; Tue, 17 Dec 91 02:18:59 -0800
Received: by fsa.cpsc.ucalgary.ca (4.1/CSd1.2)
	id AA00949; Tue, 17 Dec 91 03:17:11 MST
Return-Path: <wuth@castrov.cuc.ab.ca>
Received: from castrov.cuc.ab.ca by ajfcal.cuc.ab.ca (5.65c/Cuc2.2)
	id AA25123; Tue, 17 Dec 1991 02:05:49 -0700
Received: by castrov.cuc.ab.ca (V1.13/Amiga)
	id AA08411; Mon, 16 Dec 91 23:31:30 MST
Date: Mon, 16 Dec 91 23:31:30 MST
Message-Id: <9112170631.AA08411@castrov.cuc.ab.ca>
From: wuth@castrov.cuc.ab.ca (Brett Wuth)
To: voynich@rand.org
Subject: Re: Notes (2)
Status: OR

Robert Firth <firth@SEI.CMU.EDU> writes:

>But there are some peculiarities.  One is that the inflection scheme
>is very, very consistent, far more so than latin.  There seems to be
>only one major declension or conjugation.  Another is that what seem
>to be grammatical endings are also words in their own right:
>
>	AM	[70]	AN	[24]	AR	[58]
>
>as if latin had words "us", "um", "os" to match its endings.  That
>may have been the case in proto-indo-european, but it's not the
>case with greek, latin or most modern european languages.  It would
>also be surprising in a synthetic language devised by a european;
>such languages tend to follow the same patterns as the natural
>languages known to the devisor: as examples I cite Volapu"k and
>Esperanto.

Esperanto might be a counter example to your point.  Many afixes can
also be treated as words in their own right.  For example:

   -il-  (tool used for)  e.g. trancxilo = cut + tool = knife
    ilo = tool

   -id-  (child of)       e.g. bovido = cow + child = calf
    ido = child

   -ajx- (sustantive form of) e.g. memorajxo = memory + substantive = momento
    ajxo = thing


On another tack, it may be that these "words" are simply
suffixes with a mistaken space inserted.

Cheers,

--
Brett Wuth   (403) 242-0848
wuth@castrov.cuc.ab.ca      BCWuth@uncamult.bitnet     wuth@castrov.uucp

From gauss!Csli.Stanford.EDU!kornai Tue Dec 17 1:42:00 PST 1991
Received: by gauss; Tue Dec 17 04:47:55 EST 1991
Received: by inet.att.com; Tue Dec 17 04:47 EST 1991
Received: from Csli.Stanford.EDU by rand.org; Tue, 17 Dec 91 01:42:06 -0800
Received: by Csli.Stanford.EDU (4.1/inc-1.0)
	id AA00924; Tue, 17 Dec 91 01:42:01 PST
Date: Tue, 17 Dec 91 1:42:00 PST
From: Andras Kornai <kornai@Csli.Stanford.EDU>
To: voynich@rand.org
Subject: Some comments on Robert Firth's notes
Message-Id: <CMM.0.90.2.692962920.kornai@Csli.Stanford.EDU>
Status: OR

In his "Notes on the Voynich Manuscript - Part 2" Robert Firth writes:

>But there are some peculiarities.  One is that the inflection scheme
>is very, very consistent, far more so than latin.  There seems to be
>only one major declension or conjugation.  

Not necessarily true -- it seems to me that the method you used for
finding the root+suffix combinations can only lead to consistent
inflection patterns. Have you considered grouping "similar" suffixes
together, where "similar" is defined as having the same final and/or
medial characters (for suffixes), and as the same initial and/or medial
characters (for roots)? (For the rationale supporting this peculiar
definition of "similar" see at the end).

>Another is that what seem
>to be grammatical endings are also words in their own right:
>
>	AM	[70]	AN	[24]	AR	[58]
>
>as if latin had words "us", "um", "os" to match its endings.  That
>may have been the case in proto-indo-european, but it's not the
>case with greek, latin or most modern european languages.  It would
>also be surprising in a synthetic language devised by a european;
>such languages tend to follow the same patterns as the natural
>languages known to the devisor: as examples I cite Volapu"k and
>Esperanto.

While I agree with the reasoning I must dispute the facts.  At least
one european language, namely Hungarian (which is of course not
indo-european but for the purposes of this discussion is definitely
"european") manifests this phenomenon very clearly: case suffixes
such as bAn or nAk (the capital A sands for an archiphoneme that gets
realized as a or e depending on vowel harmony) appear as stems. When
pronouns are case-marked they appear as the case-sufix (root or stem)
followed by person/number marking (suffixes) rather than the
pronoun+case_suff that would be expected. I understand this
phenomenon is not peculiar to Hungarian (though I don't at the 
moment have access to grammars of languages (Finnish?) that would
provide similar examples).

But on the whole I totally agree with Mr. Firth that the most
reasonable working hypothesis about the Voynich ms. is that it is
plain text (perhaps heavily abbreviated, but still plain, as opposed
to encrypted) written in a natural or natural-like language in an
unusual script. Rather than Chinese brought back to Europe by Marco
Polo, the authors could be Hungarian university students -- there
were quite a few at all major medeival centers of learning. This
hypothesis is of course offered with tounge firmly in cheek, but
the larger point that languages outside the indo-eoropean family 
should also be considered, is quite serious.

As for the definition of "similar", in many languages the sounds at
the stem-suffix boundary are influenced by one another (internal
sandhi) so that we expect the last few characters in the stem and the
first few characters in the suffix to be somewhat altered.  The medial
characters in the suffix can be influenced by long-distance processes
such as vowel harmony (here I assume that both stems and suffixes fit
into a monosyllabic C*VC* pattern, where V is vowel and C* is a
consonant cluster) and the medial characters in the stem can be
affected by stress-shift (like in English plane/planar title/titular
table/tabular). In contrast to this, stem-initial and suffix-final
sounds tend to be quite stable (there are exceptions, but the overall
tendency is clear).

Andras Kornai

PS. Given the constant exhortations by John Baez that serious
participants should read D'Impero first, I was somewhat hesitant to
post to the list, especially as the Stanford library doesn't have the
book. But now that the 800 number was posted us lurkers at least have
a chance to catch up with the rest of you guys (seems like I'm talking
myself into some transcription duty here:-)! 

PPS. Even with the best of intentions it will be some weeks before 
the book arrives. In the meantime, I understand there are some 
publicly ftp-able GIF pages already available. Is there such a
thing as a gif reader for a Sparc (SunOS 4.1.1, cgthree screen)
running X-windows (X11R5)?



From gauss!SEI.CMU.EDU!firth Tue Dec 17 09:34:49 0500 1991
Received: by gauss; Tue Dec 17 10:13:01 EST 1991
Received: by inet.att.com; Tue Dec 17 10:12 EST 1991
Received: from BP.SEI.CMU.EDU by rand.org; Tue, 17 Dec 91 06:35:00 -0800
Received: by bp.sei.cmu.edu (5.64/2.5)
        id AA28735; Tue, 17 Dec 91 09:34:49 -0500
Date: Tue, 17 Dec 91 09:34:49 -0500
From: firth@SEI.CMU.EDU
Message-Id: <9112171434.AA28735@bp.sei.cmu.edu>
To: voynich@rand.org
Subject: D'Imperio
Status: OR

Just a note to say the D'Imperio monograph arrived by
inter-library loan.  It will take some time to read,
but gosh, what a disappointment the drawings are.  I
just compared the very bad photocopy of Pisces in the
front of Brumbaugh with her drawing on p 88.  What
we really need is some professional full colour 35mm
slides of the original.

Robert

From gauss!trl.OZ.AU!j.guy Tue Dec 17 10:32:46 EST 1991
Received: by gauss; Mon Dec 16 19:13:41 EST 1991
Received: by inet.att.com; Mon Dec 16 19:13 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Mon, 16 Dec 91 15:33:50 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA18776; Tue, 17 Dec 91 10:32:46 EST
Date: Tue, 17 Dec 91 10:32:46 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112162332.AA18776@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Transcription problems
Status: OR


I heartily agree with John Baez's recent comments on his attempt at
transliteration folio 93r, even with his "grumpy" comment about
my suggestion of how to distinguish between Currier's J and 7.

One of the many problems with which we are faced is that we do not
have a good copy of Currier's alphabet. The one in D'Imperio is, I
believe, all we have, and the Voynich letters in there are so very
poorly written that we really cannot tell at all what are the 
details that convinced Currier to distinguish between this and that
letter.

When I suggested the criterion I did to distinguish between J and 7,
it was only half-heartedly, for I myself saw precious little 
difference (none, really) in D'Imperio's table. I only hit upon
two obvious variants of J in folio 3v, and assumed that the second,
which is rarer, was what Currier meant by "7". Hardly something
one can reliably work on.

John Baez also writes:

"2)  I have trouble with R's versus S's"

Well, I don't  :-} (that wasn't a "smiley", that was a "smirky")
What John meant was: "I have trouble with R's versus 2's".


Here are the clues I use:

The part of the R that rests on the line is c-like. From the
thickness of the stroke, one can tell that it was written
counter-clockwise, just like we write "c".

The part of the 2 that rests on the line is Voynich i-like.
It was written from top to bottom, leaning considerably left.
The stroke is straight, and the same thickness its whole length.

The top part of R and 2 was written second, from bottom to top,
counter-clockwise, its bottom (start) linked to the top (start) of the first stroke.

Another hint is in the letter that precedes R or 2:

C *NEVER* precedes 2
A and I *NEVER* precede R

(That's according to Currier, and I wasn't aware of that particular
observation by him, when I wrote my Cryptologia article in which
I mentioned a similar phenomenon).

John Baez further remarks:

"3)  I note a tendency to be less willing to note questionable
characters, and more eager to just take a stab in the dark, as one
tires of transcribing."

Oh yes, oh yes, how true. That's why we'd need a transcription system
that so intuitive that one wouldn't be called upon making split-hair
decisions all the time. Perhaps I'm the lucky one: my reproductions
from that microfilm are so awful that I have a ready-made excuse not to do any transliterating.  :-]  (now, what was that? I'm afraid this
business of designing fonts is causing some serious overheating of the
brain cells)





From gauss!math.mit.edu!jbaez Tue Dec 17 11:09:53 EST 1991
Received: by gauss; Tue Dec 17 11:38:57 EST 1991
Received: by inet.att.com; Tue Dec 17 11:38 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Tue, 17 Dec 91 08:11:10 -0800
Received: from riesz (RIESZ.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA03548; Tue, 17 Dec 91 11:06:37 EST
Received: by riesz; Tue, 17 Dec 91 11:09:53 EST
Date: Tue, 17 Dec 91 11:09:53 EST
From: jbaez@math.mit.edu
Message-Id: <9112171609.AA01506@riesz>
To: voynich@rand.org
Status: OR

While I again recommend that all serious Voynichologists obtain
D'Imperio's book, and make it clear to Aegean Park Press that demand for
it far exceeded their wildest dreams :-), in the meantime I'll quote
section 4.4.1, "Phenomena in the Text Which Must be Accounted for by Any
Theory."  

The first 5 points are so obvious if you've seen any of the ms that I'll
skip them.  The rest are rather odder, and I am sure that any successful
explanation of the Voynich must grapple with all, or at least most, of
them.  Here let me remind you that in any good mystery story the key to
a successful solutionlies, not in shrugging off uncomfortable minutiae
that don't fit in, but in trying to find a theory that fits ALL of them.
Of course mystery novels are more neat than real life, where one expects
a few loose ends.  Still, in my view no theory has arisen which
successfully address most of these points:

6.  The same "word" is frequently repeated 2, 3, or more times in
succession.

7.  Many "words" differ from each other by only one or two symbols, and
such "words" often occur in immediate succession.

8.  Certain symbols occur characteristically at the beginnings, middles,
and ends of "words," and in certain preferred sequences.

9.  Certain symbols appear very rarely, and only on certain pages,
indicating some special function or meaning.

10.  There are very few doublets (repetition of the same letter twice in
succession), and these involve primarily the symbols C and I [here I
am using Currier notation - Baez], occasionally also 9, 8, and O.

11.  Very few symbols occur singly (as one-letter "words") in running
text; these are primarily "2" and "9". [Note, though, that 2 may really be
occuring at the end of a word here.  Do we ever see a lone 2 at the
beginning of a line? - Baez]

12.  "Prefix-like" elements are tacked in front of certain "words" that
also occur commonly without them; such prefixed elements are 4O, O, and
9.

13.  The symbol 4 occurs almost invariably followed by O, and joined to
it by an extension of the crossbar of the 4; the resulting compound
symbol is rarely seen elsewhere than at the beginning of words.

14.  On most herbal folios, the first line of the first paragraph begins
with a very small set of symols, primarily P, F, B, and V; these are
usually followed by S, Z, O, 9, AM, or 89.  No trace can be found of the
alphabeticity that would be expected if the herbal paragraphs began
with the names of plants in alphabetical order as was usual in many
early herbals.

15.  Single "words" occuring as lagels next to stars, "drug containers,"
plant sketches, or other pictorial elements in various drawings very
*rarely* begin with the 4 looped symbols [P, F, B, V]; instead, they
often start with O, 8, 9, and occaisionally 2 and S.

------
These, together with the low entropy of the text and the apparent
peculiarities of the first word of each line, and the statistical
differences between "hand A" and "hand B" (I must note I've never yet myself
seen the difference between these "hands"), make it difficult for me to
believe that the text can be explained by the presence of Hungarian
students.  :-)  Or, for that matter, any hypothesis I know except highly
sophisticated deliberate gibberish, or elaborate use of nulls (dummy text).
I would not be surprised, however, if neither of these turned out to be the
real answer.

From gauss!math.mit.edu!jbaez Tue Dec 17 11:44:00 EST 1991
Received: by gauss; Tue Dec 17 12:29:48 EST 1991
Received: by inet.att.com; Tue Dec 17 12:29 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Tue, 17 Dec 91 08:45:17 -0800
Received: from riesz (RIESZ.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA03719; Tue, 17 Dec 91 11:40:45 EST
Received: by riesz; Tue, 17 Dec 91 11:44:00 EST
Date: Tue, 17 Dec 91 11:44:00 EST
From: jbaez@math.mit.edu
Message-Id: <9112171644.AA01564@riesz>
To: voynich@rand.org
Subject:  Getting the manuscript from the Beinecke 
Status: OR

Not having heard from Robert Babcock for two 2 weeks since I wrote him --    
he's the one to contact about getting copies of the Voynich ms. -- I called the
Beinecke Rare Book Library at Yale just now.  I spoke to a woman who didn't
seem to rapidly absorb what I was after, and finally she asked me *which*
manuscript I was trying to get a copy of.  I said "The Voynich Manuscript" and
she said "OH, you'll have to talk to Robert Babcock about that, but he's out
of town and won't be back until January 2nd.  Have you all been calling her
about the Voynich or does this just go to *prove* that there is a massive
conspiracy led by a Cathari "Isis cult" that's trying to make it difficult for
people to get this manuscript?  Oh well...

jb

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Tue Dec 17 09:02 PST 1991
Received: by gauss; Tue Dec 17 13:21:13 EST 1991
Received: by inet.att.com; Tue Dec 17 13:21 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Tue, 17 Dec 91 09:11:24 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Tue, 17 Dec 91 09:16:06 PST
Date: Tue, 17 Dec 91 09:02 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: non Indo-European languages (Kornai's conjecture)
To: voynich@rand.org
Message-Id: <CC2835A8937F40669D@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

  Vowel harmony is an important feature of many Mongol and Turkic languages
too -- and the use of post-positions with (or without depending on
spelling conventions) variations due to vowel harmony could lead to a
fairly regular "suffix" structure of the sort people have remarked on
in the Voynich ms.  This might also lead to the separability of the
suffixes -- I mean, to their often being written as separate units.
  
   The mere mention of spelling conventions however is enough to
cause all sorts of worries.  Putting aside the spelling reform conjecture,
and even the early phonetic Mandarin (why not Cantonese, eh?) conjecture,
there is still the possibility that the script represents an attempt to
transcribe a language that either has never otherwise been transcribed,
or which we know with an entirely different set of spelling conventions.

  I have seen transcriptions of Louisiana French/Creole folktales in
a spelling that is far from the usual French orthography;  it takes 
a bit of sounding-out to get the hang of it, but it seems to work
well enough.  Perhaps some cryptologist would comment on how easy it
would be to go from a simple encipherment of this folklorist's romanization
of a French dialect to a plaintext, if one didn't have any idea at the
beginning what language it might be?  When "ti" represents "petit," "gro"
represents "gros," and for all I know "farush" might represent "farouche"?
Not to mention what sort of havoc might arise from doing the same with
say Polish or Russian (fortunately ruled out by the data about regularities
in endings!).  It might well be easy -- if one had the phonemic frequency
data.  But if it were an otherwise undocumented dialect?

 I have some memory of a piece on Hildegard of Bingen's "literas ignotas"
in Pitra's _Analecta Hymnica_ -- I don't remember though whether they were
only odd letter forms or whether they were used to transcribe some 
dialect or other.  At any rate, such things do seem to have been going
on...

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Tue Dec 17 09:20 PST 1991
Received: by gauss; Tue Dec 17 13:05:55 EST 1991
Received: by inet.att.com; Tue Dec 17 13:05 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Tue, 17 Dec 91 09:22:17 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Tue, 17 Dec 91 09:26:57 PST
Date: Tue, 17 Dec 91 09:20 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: we need full-color 35 mm photographs
To: voynich@rand.org
Message-Id: <CC25A2376FFF40669D@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

How's this for a grandiose scheme?  Lobby Aegean Park Press to do a
full-color facsimile edition of the Voynich ms (with a black-and-white
"working facsimile" to accompany it and also sold separately).  It's
just about time to begin planning next year's Christmas line anyway.

It should appeal to the same market that buys other manuscript facsimiles.
It could be accompanied by a small explanatory volume with essays on
paleography, manuscript style, and decipherment speculations so far,
with perhaps historical side-notes.  More of a contribution to polite
letters, perhaps, than to scholarship, but just the thing for the
coffee table.


From gauss!rand.org!jim%mycroft Tue Dec 17 10:19:47 PST 1991
Received: by gauss; Tue Dec 17 13:43:04 EST 1991
Received: by inet.att.com; Tue Dec 17 13:42 EST 1991
Received: from mycroft.rand.org by rand.org; Tue, 17 Dec 91 10:19:50 -0800
Received: from localhost by mycroft.rand.org; Tue, 17 Dec 91 10:19:47 PST
Message-Id: <9112171819.AA01605@mycroft.rand.org>
To: voynich@rand.org
Subject: Folio 79v GIF posted
Date: Tue, 17 Dec 91 10:19:47 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

The file f79v.gif in the ftp directory is a scanned image of the plate from
Poundstone.  I'm not happy with either the size of the file or the quality
attained -- but there's a lot more text there than with f3v.gif .  I think
this is probably the other end of the spectrum in terms of how many bits we
need for the image.

Jim

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Tue Dec 17 10:36 PST 1991
Received: by gauss; Tue Dec 17 14:22:42 EST 1991
Received: by inet.att.com; Tue Dec 17 14:22 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Tue, 17 Dec 91 10:45:30 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Tue, 17 Dec 91 10:50:12 PST
Date: Tue, 17 Dec 91 10:36 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Getting the MS from the Bienecke
To: voynich@rand.org
Message-Id: <CC1B15B0E75F405A9A@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

Maybe Umberto Eco is working on another novel -- and someone has found out...

From gauss!gauss.att.com!reeds Tue Dec 17 16:26:07 EST 1991
Received: by gauss; Tue Dec 17 16:36:53 EST 1991
Received: by inet.att.com; Tue Dec 17 16:36 EST 1991
Received: from research.att.com by rand.org; Tue, 17 Dec 91 13:27:19 -0800
Message-Id: <9112172127.AA16689@rand.org>
From: reeds@gauss.att.com
Date: Tue, 17 Dec 91 16:26:07 EST
To: voynich@rand.org
Status: OR

A colleague of mine, who has been looking pretty hard at
a few pages of the MS says, that to her eye, the differences
between the accents or plumes of Currier's Z are as great
(or greater than) the differences between Currier's R and 2.

Thus, if R and 2 are distinct letters, then the several forms
of Z should be distinct letters, too.

Jim Reeds

From gauss!trl.OZ.AU!j.guy Wed Dec 18 11:33:04 EST 1991
Received: by gauss; Tue Dec 17 19:45:41 EST 1991
Received: by inet.att.com; Tue Dec 17 19:45 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Tue, 17 Dec 91 16:33:26 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA20347; Wed, 18 Dec 91 11:33:04 EST
Date: Wed, 18 Dec 91 11:33:04 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112180033.AA20347@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: several Z'z?
Status: OR


Jim Reeds writes:


A colleague of mine, who has been looking pretty hard at
a few pages of the MS says, that to her eye, the differences
between the accents or plumes of Currier's Z are as great
(or greater than) the differences between Currier's R and 2.


It had occurred to me too, a long time ago, when I first saw
reproductions, but I shut up about it because perhaps it was
my imagination, and things being hard enough as they stand...

Now, by what features does she distinguish those plumes?

From gauss!math.mit.edu!jbaez Wed Dec 18 22:10:26 EST 1991 remote from alice
Received: by gauss; Wed Dec 18 22:49:57 EST 1991
Received: by inet.att.com; Wed Dec 18 22:49 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Wed, 18 Dec 91 19:09:48 -0800
Received: from cartan (CARTAN.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA09734; Wed, 18 Dec 91 22:05:08 EST
Received: by cartan; Wed, 18 Dec 91 22:10:26 EST
Date: Wed, 18 Dec 91 22:10:26 EST
From: jbaez@math.mit.edu
Message-Id: <9112190310.AA01579@cartan>
To: voynich@rand.org
Subject: interesting books
Status: OR

Here some interesting book titles... I haven't been able to see the
books yet since they're secluded in Harvard libraries, and my Harvard
connection isn't back in town yet.

John H Brigadier Tiltman, "The Voynich Ms.: The Most Mysterious Ms. in 
the World"  --- this appears to be in the horticultural library!

Donald Laycock, "Complete Enochian Dictionary"

"Private Diary of Dr. John Dee and the Catalog of his Library of Manuscripts"

If anyone can get ahold of these and post a short report that'd be nice.

jb

From gauss!cca.ucsf.EDU!wet!naga Thu Dec 19 18:55 PST 1991 remote from alice
Received: by gauss; Thu Dec 19 22:09:48 EST 1991
Received: by inet.att.com; Thu Dec 19 22:09 EST 1991
Received: from cgl.ucsf.EDU by rand.org; Thu, 19 Dec 91 18:58:12 -0800
Received: from cca.ucsf.EDU by cgl.ucsf.EDU (5.65/GSC4.21)
	id AA17358 for voynich@rand.org; Thu, 19 Dec 91 18:57:58 -0800
Received: from wet.UUCP by cca.ucsf.EDU (5.61/GSC4.21)
	id AA29294; Thu, 19 Dec 91 18:57:42 -0800
Received: by wet.uucp (/\=-/\ Smail3.1.18.1 #18.2)
	id <m0kuaOS-0009GoC@wet.uucp>; Thu, 19 Dec 91 18:55 PST
Message-Id: <m0kuaOS-0009GoC@wet.uucp>
Date: Thu, 19 Dec 91 18:55 PST
From: wet!naga@cca.ucsf.EDU (Peter Davidson)
To: voynich@rand.org
Subject: McKenna/Levitov
Status: OR

 
Terence McKenna has been researching the Voynich ms. for some years.
He published a discussion of the problem and a review of Levitov's
book in an article in Gnosis, No. 7 (Spring 1988), entitled "Has the
World's Most Mysterious Manuscript Been Read at Last?".  Among his
tapes on various subjects is one entitled "The Voynich Manuscript"
which is probably available either from Dolphin Tapes, POB 71, Big Sur,
CA 93920, or from  Sound Photosynthesis, POB 2111, Mill Valley, CA 94942.
 
Since this group has been formed to find a solution to the problem of
what the Voynich ms. is and is about, the implication is that Levitov's
proposed solution (that the ms. is the only surviving primary document
of the Cathar faith and that it is a highly polyglot form of Medieval
Flemish with a large number of Old French and Old High German loan
words, written in a special script) is not accepted.  If not, why not?
 
Could someone contribute a summary of Levitov's claims and reasons,
if any, why they don't stand up to criticism (if in fact they don't).
 
 

From gauss!rand.org!jim%mycroft Thu Dec 19 22:19:19 PST 1991 remote from alice
Received: by gauss; Fri Dec 20 01:29:20 EST 1991
Received: by inet.att.com; Fri Dec 20 01:29 EST 1991
Received: from mycroft.rand.org by rand.org; Thu, 19 Dec 91 22:19:22 -0800
Received: from localhost by mycroft.rand.org; Thu, 19 Dec 91 22:19:19 PST
Message-Id: <9112200619.AA06919@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: McKenna/Levitov 
In-Reply-To: Your message of Thu, 19 Dec 91 18:55:00 -0800.
             <m0kuaOS-0009GoC@wet.uucp> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Thu, 19 Dec 91 22:19:19 PST
Sender: jim%mycroft@rand.org
Status: OR

> Could someone contribute a summary of Levitov's claims and reasons,
> if any, why they don't stand up to criticism (if in fact they don't).

I replied privately to this pointing at Guy's review in the file
rand.org:pub/jim/levitov -- but this deserves a wider question, i.e.
what's wrong with all the other claims?  We should have a collection of
the definitive critiques of each, including Newbold, Feely, Strong (to
the extent that we can comment on it in the absence of his actual decryption
method), and Brumbaugh.

I pressed Strong for details in a few letters before his death, but he was
so offended by Friedman's critique and Kahn's disinterest that he decided
never to publish it.

Jim Gillogly

From gauss!cca.ucsf.EDU!wet!naga Fri Dec 20 02:57 PST 1991 remote from alice
Received: by gauss; Fri Dec 20 06:47:58 EST 1991
Received: by inet.att.com; Fri Dec 20 06:47 EST 1991
Received: from cgl.ucsf.EDU by rand.org; Fri, 20 Dec 91 03:31:17 -0800
Received: from cca.ucsf.EDU by cgl.ucsf.EDU (5.65/GSC4.21)
	id AA20493 for voynich@rand.org; Fri, 20 Dec 91 03:31:06 -0800
Received: from wet.UUCP by cca.ucsf.EDU (5.61/GSC4.21)
	id AA03171; Fri, 20 Dec 91 03:30:58 -0800
Received: by wet.uucp (/\=-/\ Smail3.1.18.1 #18.2)
	id <m0kuhvK-000CqwC@wet.uucp>; Fri, 20 Dec 91 02:57 PST
Message-Id: <m0kuhvK-000CqwC@wet.uucp>
Date: Fri, 20 Dec 91 02:57 PST
From: wet!naga@cca.ucsf.EDU (Peter Davidson)
To: voynich@rand.org
Subject: How old is the MS?
Status: OR

 
>From the McKenna article in Gnosis, No. 7: " The manuscript has never
been physically analyzed, which would settle, once and for all, at least
the century of its origin.  ...  Levitov does not mention the physical
manuscript.  Yet it seems obvious that one of the first steps that
should be taken would be to attempt to confirm the 13th century origin
date for the manuscript.  ...  Surely it should be possible to determine
whether the manuscript was written in the 13th or the 16th century."
 
If it was the 16th century then Dee and Kelley become the prime suspects
for the "two hands", but if it was the 13th century then we can rule
out McKenna's hypothesis that it was a forgery produced to palm off to
the Emperor Rudolph for a princely sum.
 

From gauss!math.mit.edu!jbaez Fri Dec 20 13:39:09 EST 1991 remote from alice
Received: by gauss; Fri Dec 20 14:22:48 EST 1991
Received: by inet.att.com; Fri Dec 20 14:22 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Fri, 20 Dec 91 10:42:00 -0800
Received: from cayley (CAYLEY.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA15059; Fri, 20 Dec 91 13:37:07 EST
Received: by cayley; Fri, 20 Dec 91 13:39:09 EST
Date: Fri, 20 Dec 91 13:39:09 EST
From: jbaez@math.mit.edu
Message-Id: <9112201839.AA08058@cayley>
To: voynich@rand.org
Subject: How old is the MS?
Status: OR


As discussed earlier in this mailing list, carbon dating should permit
us to tell the difference between a 16th-century and a 13th-century origin
of the vellum on which the Voynich was written.  The key here is to A) get
the Beinecke to permit it, and B) get a laboratory to do it for free.  (It
would cost ~$1000 otherwise.)  Nate and I have some leads on B), but for A)
it would be good if we could get an expert on medieval manuscripts -- preferably
one from Yale -- to back our cause.  Please, everyone, try to pull strings
with your friends and friend-of-friends, and get a well-known scholar interested!

jb

From gauss!cl.cam.ac.uk!Michael.Roe Sat Dec 21 15:15:29 +0000 1991 remote from alice
Received: by gauss; Sat Dec 21 10:49:01 EST 1991
Received: by inet.att.com; Sat Dec 21 10:48 EST 1991
Received: from swan.cl.cam.ac.uk by rand.org; Sat, 21 Dec 91 07:15:49 -0800
Received: from faeroes.cl.cam.ac.uk by swan.cl.cam.ac.uk 
          with SMTP (PP-5.7) to cl id <25429-0@swan.cl.cam.ac.uk>;
          Sat, 21 Dec 1991 15:15:35 +0000
Received: by uk.ac.cam.cl.lion (15.11/SMI-3.0DEV3) id AA28442;
          Sat, 21 Dec 91 15:15:31 gmt
Message-Id: <9112211515.AA28442@uk.ac.cam.cl.lion>
To: jbaez@math.mit.edu
Cc: voynich@rand.org
Subject: Re: How old is the MS?
In-Reply-To: Your message of Fri, 20 Dec 91 13:39:09 EST . <9112201839.AA08058@cayley>
Date: Sat, 21 Dec 91 15:15:29 +0000
From: Mike Roe <Michael.Roe@cl.cam.ac.uk>
Status: OR

  
> As discussed earlier in this mailing list, carbon dating should permit
> us to tell the difference between a 16th-century and a 13th-century origin
> of the vellum on which the Voynich was written.  

I believe that there is already fairly strong evidence from the style of the 
illustrations and calligraphy that the MS is not much older than the 
15th-century. 
For example, in [Brumbaugh76], the following evidence is given:
- The cipher table includes J, V, and W [presumably he means the one on f1r].
- There is a two-handed clock on f85
- The costume of the medallion on the Sagittarus map
- The Arabic numerals in the margin of f49r.

In [ONeill4], some of the botanical illustrations are identified as plants
from the New World, suggesting that it was written after Columbus. 
(Although it is possible that some imaginary plants draw in the 13th century
might happen to look like real plants which were discovered later).

The style of the nudes in for example f79v is also evidence against a 13th
century origin.

In view of this, I don't think it would be worthwhile to radiocarbon date
the MS: It is quite clear that it was written later than the 13th-century, and
carbon dating will provide little extra information.

There are other techniques which are better at discriminating between more
recent dates. For example, it might be possible to analyse the composition
of the inks. (This would reveal if modern inks were used. A forger could
have deliberately used old ink recipies, but might not have done). In
Barlow's review of [dImperio78], a Dr. Carter is quoted as saying that
one of the inks is ``just like an ordinary red ink of today''.  This might be
worth following up.

Michael Roe
Cambridge University Computer Laboratory\

References
==========

@article{Barlow86,
author = "Barlow, Michael",
title = "The Voynich Manuscript --- By Voynich?",
journal = "Cryptologia",
volume = "10",
number = "4",
pages = "p210 -- 216",
month = "October",
year = 1986}

@article{Brumbaugh76,
author = "Brumbaugh, Robert S.",
title = "The Voynich ``Roger Bacon'' Cipher Manuscript : Deciphered Maps of the
Stars",
journal = "Journal of the Warburg and Cortauld Institutes",
volume = 39,
pages = "p139",
year = 1976}

@article{ONeill44,
author = "O'Neill, Hugh",
title = "Botanical Remarks on the Voynich MS",
journal = "Speculum XIX",
pages = "p.126",
year = 1944}

From gauss!rand.org!jim%mycroft Wed Dec 25 06:44:25 PST 1991 remote from alice
Received: by gauss; Wed Dec 25 10:09:04 EST 1991
Received: by inet.att.com; Wed Dec 25 10:09 EST 1991
Received: from mycroft.rand.org by rand.org; Wed, 25 Dec 91 06:44:27 -0800
Received: from localhost by mycroft.rand.org; Wed, 25 Dec 91 06:44:25 PST
Message-Id: <9112251444.AA12751@mycroft.rand.org>
To: voynich@rand.org
Subject: VMs - letter from Prescott Currier
Date: Wed, 25 Dec 91 06:44:25 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

[I sent Capt. Currier a transcript of our conversation through
the first part of 4 Dec 91, and asked about the trouble we're having
distinguishing between 7 and J.  I will be sending a copy of his original
transcription alphabet to everybody who is doing volunteer transcription
work.  Let me know if you also want a copy of his paper -- I don't want to
send it to everybody, because of the copying and postage. - JJG]

---------------------------------------------------------------------------

							 Damariscotta, Me.
							  16 Dec. 1991

Dear Jim -

        I must say that I was both surprised and pleased to receve your
recent letter.  I had no idea that anyone was still interested in the
Voynich Manuscript.  I have now read through your computer conversations
and am struck by the character and qualifications of your group.  I feel
that if any effort toward a solution of the VMs mystery can succeed yours
has the best chance yet.  As you are perhaps aware I hold in low regard
most of the work that has been done on the manuscript over the past few
decades for the reason that most of it seems to lack scientific
underpinning.  In my view little substantive can be accomplished without
due consideration of the conclusions and observations set forth in my 1976
paper.  I am sure you are familiar with its content since reference to it
is made in the material you sent me.  However I'm sending you another copy
just in case; also a copy of the original transcription alphabet.

	In answer to your one specific question about the difference between
J and 7 -  J [drawing that looks like J on the alphabet chart in D'Imperio]
is the first of the finals series [drawings for JKL on chart], while 7
[drawing that looks like the Krischer letter below [Currier] F on the chart]
is one of a pair of rare, ampsersand-like characters, the second being
[drawing that looks like 6 on the chart].  The first seems to be composed
of [drawing of Currier I] plus [drawing  of right half of J], the second
being [drawing of C] plus [drawing of right half of J].  [Note, JJG: both
ampersands are closed at the bottom -- the 7 drawing in D'I should be
considered inoperative: increase the loop, and angle the lower left straight
piece over to close with the bottom of the right stroke.]

	My only immediate comment on the various opinions and suggestins
appearing in your CompCons (with which I am in general agreement), is to
explain my reason for adopting specific symbols for "end of Word" "end of
line" and "end of para".   My study of the VMs led me to believe that these
positions were somehow functional and I wanted to be able to sort on any
of them.

	Just in case you're interested, my own qualifications include an
original AB in Romance languages, Japanese in WWII, graduate of Navy Russian
language school and an advanced degree in Comparative Philology
(Indo-european) from the Univ. of London.  Plus work on Linear B, Hittite
etc.  While at the Univ. of London I had a reader's card at the British
Museum where I spent quite [a lot] of time looking at the 14th and 15th
century Herbals and other manuscripts trying to locate the origin of some
at least of the symbols in the VMs alphabet.  No luck of course.

	Unfortunately I am computer-illiterate but at 80 I think I'll leave
this to people like you.  I hope that you will see fit to keep me informed
on your progress and will seek my help whenever you feel thta I have
something to offer.  Incidentally I have what is probably a 5th or 6th
generation Xerox copy of most of the manuscript made about 25 years ago.  It
is reasonably legible and is what I used in all of my work on the VMs.
Properly filtered, enhanced and reproduced it could prove to be a fairly
decent copy.  Not to compare of course with a properly made copy of the
original Voynich at the Beinicke.

					Best regards,
				    (s) Prescott Currier

From gauss!gauss.att.com!reeds Fri Dec 27 01:03:39 EST 1991 remote from alice
Received: by gauss; Fri Dec 27 01:29:59 EST 1991
Received: by inet.att.com; Fri Dec 27 01:29 EST 1991
Received: from research.att.com by rand.org; Thu, 26 Dec 91 22:03:48 -0800
Message-Id: <9112270603.AA09532@rand.org>
From: reeds@gauss.att.com
Date: Fri, 27 Dec 91 01:03:39 EST
To: voynich@rand.org
Status: OR

1. Jacques writes

> THE DREADED "7"
> 
> I still could not figure out what the dreaded "7" looked like, from the
> description in Prescott Currier's letter to Jim Reeds, so I didn't alter my
> perfectly arbitrary rendition:
> 7=i-g
> 
> At any rate now, with TRANSLIT and Harald's FE, we can fix it easily when we know.
>

Let me mention:  it was Jim Gillogly, not I (Jim Reeds) who wrote 
to Currier.  I am in a state of diminished doubt about J and 7, but 
appreciate learning that 7 has a closed bottom.

Let me also add, that I hope Jacques' TRANSLIT program is given in source 
language form, not as an executable binary.

2.  How I spent my Christmas vacation:  after transcribing my quota of
folios, (not yet entered into the computer) I snooped around in occult 
bookstores and bought (and dipped into) a couple of Dee books.  The most 
generally useful one seems to be

Peter French,
John Dee, the World of an Elizabethan Magus
Routledge, 1972?, reprinted by Dorset Press, 1989, as ISBN 0-88029-445-0

Another (more recent and high scholarly) by Clulee, is being mailed to
me, so I cannot cite full title, etc.  It seems to be an expanded version
of Clulee's U of Chicago thesis under Alan Debus.

There exist modern editions of Trithemius's Steganographia (trans.
by Fiona Tait ansd Christopher Upton, Edinburgh: Magnum Opus Hermetic
Sourceworks, 1982) and of Casaubon's True and Faithful Relation.  
(Askin, 1974).  These books, like Laycock's Complete Dictionary,
are prized by occult book freaks, and the clerk at Shambhala Books 
in Berkeley laughed at me when I asked if he had them in stock.

It seems clear (from French, Clulee, D'Imperio) that if you want to
understand what Dee DID you should read Casaubon, even though he
(according to the Shambhala clerk) is unreliable.  (Peter French
pp.12-13 tells an interesting story about WHY Casaubon, who seems a
sort of 17-th century Amazing Randi, wrote his book.)

French lists the BM's Sloane MS 3189, the ``Liber mysteriorum
sextus et sanctus'' as being in Kelley's hand, in case we want to
look for handwriting similarities.  French gives samples of Dee's
handwriting, both plain and elegant (in a letter to the Queen).
It is very legible and doesn't look anything like the VMS.

He also gives (in plate 10) the transcription of the first seance
Kelley and Dee had together, Sat 10 March 1582, the day K and D first
met.  Very roughly, the dialoge went like this.

DEE:  Are you an angel callable on this stone?  

URIEL: Yup.

DEE: Are there any more beside you?

URIEL: Michael and Raphael.  But Michael est priceps in operibus ***

DEE:  Is my book of SOYGA of any exellency?

URIEL:  God's good angels gave it to Adam.

DEE:  Will you give me any instructions, so that I may read those Tables of
Soyga?

URIEL:   I can, but solus Michael illius Libri est interpriator.

DEE:  I was told that after I could read that boke,  I should Live but
two years and a half.

URIEL: Thou shalt live an hundred and od (?) years.

DEE:  What's Michael's email address?

URIEL:  michael@up.there

DEE:  Oh, my great and long desire hath been to be able to read those 
Tables of Soyga!

...
Note the sequence:  1, establish credentials, 2, find out if they can
read the book of Soyga, 3, find out if its safe to read, 4, find out
how to reach Michael.

IF the Book of Soyga is the Voynich MS, then I think Dee qualifies as 
a charter member of our group!  (If we may ask for help from retired NSA
types, Dee was entitled to ask for angelic help, no?)

And given the timing (Dee met Kelley that day, but Dee has long desired 
a construe), Kelley would not be the forger.  That is, the book of SOYGA
is not the book of St Dunstan, and hence at most one of them is the VMS.

3.  Another line of investigation, best left to experts.  There was vigorous
production of illustrated printed herbals in the 16th century, whose pictures
were copied from each other with a freedom which is breathtaking to moderns'
sensibilities.  (See, for example, K. Reeds's forthcoming Botany in
Medieval and Renaissance Universities (Garland, 1992))  It might turn out
that the herbal or pharmaceutical pictures in the VMS were copied in whole
or in part from some standard illustrated herbal.  This does not mean 
identifying the plants: maybe only some kind of detail is copied, or just
layout of picture is copied.  Copying from an herbal seems quite likely 
to me if the VMS is forged.

Jim Reeds


From gauss!cca.ucsf.EDU!wet!naga Fri Dec 27 06:37 PST 1991
Received: by gauss; Fri Dec 27 20:12:26 EST 1991
Received: by inet.att.com; Fri Dec 27 20:12 EST 1991
Received: from cgl.ucsf.EDU by rand.org; Fri, 27 Dec 91 17:00:56 -0800
Received: from cca.ucsf.EDU by cgl.ucsf.EDU (5.65/GSC4.21)
	id AA24593 for voynich@rand.org; Fri, 27 Dec 91 17:00:51 -0800
Received: from wet.UUCP by cca.ucsf.EDU (5.61/GSC4.21)
	id AA22271; Fri, 27 Dec 91 17:00:11 -0800
Received: by wet.uucp (/\=-/\ Smail3.1.18.1 #18.2)
	id <m0kxIgU-000GcmC@wet.uucp>; Fri, 27 Dec 91 06:37 PST
Message-Id: <m0kxIgU-000GcmC@wet.uucp>
Date: Fri, 27 Dec 91 06:37 PST
From: wet!naga@cca.ucsf.EDU (Peter Davidson)
To: voynich@rand.org
Subject: 1582-88
Status: OR

 
Jim Reeds writes:
>He [French] also gives (in plate 10) the transcription of the first seance
>Kelley and Dee had together, Sat 10 March 1582, the day K and D first met.
 
It's worth noting (this from Kent's chapter in the Brumbaugh book) that:
 
(i)  Dee met Francis Bacon for the first time in August 1582 and
presumably showed him his Roger Bacon manuscripts.
 
(ii)  There was "a catalogue of Dee's library, prepared in 1583, just
before a mob at Mortlake destroyed many of his books ..."  Is this
catalogue extant?  If so, does it list anything which might be the VMS?
 
(iii)  Dee visited Prague several times during 1584-88, apparently the
only period during which he could have passed the MS to the Emperor Rudolf.
 
P.S.  I hope Jim G. will make Jacques G.'s program available to those
of us who don't have FTP ability.  Could it be uuencoded and mailed?
 

From gauss!trl.OZ.AU!j.guy Fri Dec 27 12:18:16 EST 1991 remote from alice
Received: by gauss; Thu Dec 26 20:50:17 EST 1991
Received: by inet.att.com; Thu Dec 26 20:50 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Thu, 26 Dec 91 17:18:34 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA28422; Fri, 27 Dec 91 12:18:16 EST
Date: Fri, 27 Dec 91 12:18:16 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112270118.AA28422@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: a utility and some errata
Status: OR


A couple of hours ago I sent to jim@rand.org a utility for putting in the pub/jim directory.
It consists of a program for translating from any (sensible) transcription system into any
other (sensible) system, and back again. The program is called TRANSLIT and works like this.
Say you have the file called VOYNICH (D'Imperio's original transcription in Currier's system)
and want to translate it into Bennett's system, with the result in a new file called 
VOYNICH.BEN. Just say:

TRANSLIT VOYNICH DIMP2BEN VOYNICH.BEN

DIMP2BEN is a short file containing a list of equivalences between Currier's and Bennett's
transliteration systems.

To change from Bennett's into Currier's you say:

TRANSLIT VOYNICH.BEN -DIMP2BEN VOYNICH.TMP

(the minus sign in front of DIMP2BEN tells TRANSLIT to translate in the other direction)

To avoid catastrophes, I have written TRANSLIT so that it will absolutely refuse to overwrite
an existing file.

I have provided 3 equivalence files with TRANSLIT:

DIMP2BEN   Currier to Bennett, as used in D'Imperio's original file (VOYNICH)
CURR2BEN   Currier to Bennett: the same, except that the word separator is a period (.)
           instead of a slash (/). That is the transcription in VOYNICH.NOW

CURR2GUY   Currier to the system I suggested recently.

Now I made two stupid mistakes in file CURR2GUY. On line 4 I wrote:
2=2
that should have been:
2=z

On line 6 I wrote:
R=z
that should have been:
R=2

(I'll send the corrected stuff to jim@rand.org shortly.


Corrigenda in VOYNICH and VOYNICH.NOW

I have no decent reproductions to speak of, but I'm sure I have spotted some mistakes here:

Line <00104A>  $OM  that must be 4OM I suppose
Line <3v.4>    408AR  that 0 (zero) is really an O 
Line <3v.13>   40FSA  ditto
Line <04401A>  O8AIIIE9 ought to be O819 (since Currier's 1 is made up of three I's and one E)

THE DREADED "7"

I still could not figure out what the dreaded "7" looked like, from the
description in Prescott Currier's letter to Jim Reeds, so I didn't alter my
perfectly arbitrary rendition:
7=i-g

At any rate now, with TRANSLIT and Harald's FE, we can fix it easily when we know.



From gauss!rand.org!jim%mycroft Fri Dec 27 17:18:46 PST 1991
Received: by gauss; Fri Dec 27 20:51:07 EST 1991
Received: by inet.att.com; Fri Dec 27 20:51 EST 1991
Received: from mycroft.rand.org by rand.org; Fri, 27 Dec 91 17:18:50 -0800
Received: from localhost by mycroft.rand.org; Fri, 27 Dec 91 17:18:47 PST
Message-Id: <9112280118.AA15176@mycroft.rand.org>
To: voynich@rand.org
Subject: Re: 1582-88 
In-Reply-To: Your message of Fri, 27 Dec 91 06:37:00 -0800.
             <m0kxIgU-000GcmC@wet.uucp> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Fri, 27 Dec 91 17:18:46 PST
Sender: jim%mycroft@rand.org
Status: OR


> wet!naga@cca.ucsf.EDU (Peter Davidson) writes:
> 
> P.S.  I hope Jim G. will make Jacques G.'s program available to those
> of us who don't have FTP ability.  Could it be uuencoded and mailed?

It could be -- that's how Jacques sent it to me.  However, I'd rather
not get into the business of doing special mailings.  Anybody know how
much trouble it is to set up a mailserv system on a vax?  Anybody want
to volunteer to process individual file transfer requests for uucp and
bitnet members?

	Jim Gillogly

From gauss!gauss.att.com!reeds Fri Dec 27 20:39:35 EST 1991
Received: by gauss; Fri Dec 27 21:02:27 EST 1991
Received: by inet.att.com; Fri Dec 27 21:02 EST 1991
Received: from research.att.com by rand.org; Fri, 27 Dec 91 17:39:44 -0800
Message-Id: <9112280139.AA19929@rand.org>
From: reeds@gauss.att.com
Date: Fri, 27 Dec 91 20:39:35 EST
To: voynich@rand.org
Status: OR

Peter Davidson writes
>  
> It's worth noting (this from Kent's chapter in the Brumbaugh book) that:
> [...]  
> (ii)  There was "a catalogue of Dee's library, prepared in 1583, just
> before a mob at Mortlake destroyed many of his books ..."  Is this
> catalogue extant?  If so, does it list anything which might be the VMS?

French (p 212) lists British Museum (now Library): 

Harleian MS. 1879, arts.  1, 5, and 6.  `Catalogus codd. MSS. numero plus
minus 230, iam olim ut videtur, in Biblioteca Joannis Dee M. D.
conservatorum'; catalogue of Dee's printed books; catalogue of Dee's
manuscripts.  The last two items are dated 6 September 1583 and are in Dee's
handwriting.

Clulee spends a lot of time discussing these catalogues.  I think they
have been recently published, but without my Clulee in hand, I can't
give references.

PS.  Kent wrote (Brumbaugh, p 25, line 7) ``Dee had no creative power...''
which is difficult for me to accept.

From gauss!gauss.att.com!reeds Sat Dec 28 00:21:25 EST 1991
Received: by gauss; Sat Dec 28 00:49:37 EST 1991
Received: by inet.att.com; Sat Dec 28 00:49 EST 1991
Received: from research.att.com by rand.org; Fri, 27 Dec 91 21:21:34 -0800
Message-Id: <9112280521.AA21619@rand.org>
From: reeds@gauss.att.com
Date: Sat, 28 Dec 91 00:21:25 EST
To: voynich@rand.org
Status: OR

Jim Gillogly wisely writes, in response to Davidson's request
for an email version of an FTP-able file, that he'd 
>                                                                rather
> not get into the business of doing special mailings.  Anybody know how
> much trouble it is to set up a mailserv system on a vax?  Anybody want
> to volunteer to process individual file transfer requests for uucp and
> bitnet members?

I foolishly volunteer to send special mailings, until it gets too
irksome for me.  I reserve the right to get confused, to delay, to
stop the service, etc.  Please don't send me requests if you can get 
them by FTP.

Jim Reeds
reeds@research.att.com

Appendix A: files in rand.org:pub/jim 

-rw-r--r--  1 25         257583 Dec 27 17:20 digest
-rw-r--r--  1 25           9692 Dec 27 09:07 translit.zip
-rw-r--r--  1 25          10265 Dec 17 11:04 foliation
-rw-r--r--  1 25           1407 Dec 17 10:16 ReadMe.dir
-rw-r--r--  1 25          96011 Dec 17 10:15 f79v.gif
-rw-r--r--  1 25           4879 Dec 17 09:41 voyfont.troff.sh
-rw-r--r--  1 25          14088 Dec 17 09:40 voytypo.sh
-rw-r--r--  1 25          20552 Dec 17 09:31 voyfont.ps.sh
-rw-r--r--  1 25           4809 Dec 16 19:39 cur2frog.zip
-rw-r--r--  1 25           6365 Dec 15 13:29 fonts.zip
-rw-r--r--  1 25         111316 Dec 15 12:39 voynich.now
-rw-r--r--  1 25           2272 Dec 15 12:39 Log
-rw-r--r--  1 25          41754 Dec 15 11:16 fonted.zip
-rw-r--r--  1 25           3952 Dec 12 05:49 uncmt.exe
-rw-r--r--  1 25          38445 Dec 11 14:37 monkey.zip
-rw-r--r--  1 25          18726 Dec 10 18:52 cognate.zip
-rw-r--r--  1 25          18845 Dec 10 18:49 levitov
-rw-r--r--  1 25          35173 Dec  9 18:48 f3v.gif
-rw-r--r--  1 25          12572 Dec  7 20:10 guyfreq.pas
-rw-r--r--  1 25           2532 Dec  7 17:25 Welcome
-rw-r--r--  1 25           7431 Dec  7 17:11 currier
-rw-r--r--  1 25          13484 Dec  7 17:00 biblio
-rw-r--r--  1 25          40813 Nov 18 19:51 voynich.tar.Z


Appendix B: contents of rand.org:pub/jim/ReadMe.dir

	Manifest of files in this directory

Log              Changes from original D'Imperio transcriptions
Welcome          Welcome to the Voynich Manuscript mailing list
biblio           Bibliography of especially relevant articles
cognate.zip      Determine whether words in two languages are cognates (PC)
cur2frog.zip     Convert Currier notagion to Frogguy, by Guy (PC)
currier          Description of Currier notation for Voynich text
digest           Digest of messages since 4 Dec 91
f3v.gif          GIF image of folio 3 verso -- text part only
f79v.gif         GIF image of folio 79 verso, scanned from Poundstone
foliation        List of folios in Voynich Manuscript with locations
fonted.zip       Voynich font and transcription for PC, by Guy
fonts.zip        More fonts for PC, by Guy
guyfreq.pas      Program used to produces statistics in Guy Cryptologia paper
levitov          Analysis of Levitov solution claim, by Guy
monkey.zip       Program to determine entropy of text, by Guy (PC)
uncmt.exe        Program to remove comments from transcribed text, by Guy (PC)
voyfont.ps.sh    Postscript Voynich font for Currier notation, by Reeds
voyfont.troff.sh Troff Voynich font for Currier notation, by Reeds
voynich.now      Current working version of Voynich transcription
voynich.tar.Z    Original D'Imperio transcription
voytypo.sh       Voynich typo finder, Voynich parser code in C, by Reeds


Appendix C: 'vrfy voynich' output from 'telnet rand.org 25'

250-Jim Gillogly <jim@mycroft>
250-<reeds@gauss.att.com (Jim Reeds)>
250-<j.guy@trl.OZ.AU (Jacques Guy)>
250-<jbaez@math.mit.edu (John Baez)>
250-<Karl.Kluge@G.GP.CS.CMU.EDU>
250-<brill@unagi.cis.upenn.edu (Eric Brill)>
250-<A512JANW%HASARA11@pucc.PRINCETON.EDU (Jan Wim Wesselius)>
250-<em21+@andrew.cmu.edu (Eric Edward Moore)>
250-<thorin@wpi.WPI.EDU (Richard John Yanco)>
250-<yanek@mthvax.cs.miami.edu (Yanek Martinson)>
250-<sai@kauri.vuw.ac.nz (Simon McAuliffe)>
250-<kingj@hpcc01.corp.hp.com (John King)>
250-<pvs@anhep2.hep.anl.gov (Paul Schoessow)>
250-<EVANS@BINAH.CC.BRANDEIS.EDU (Ronald Hale-Evans)>
250-<nate@im.lcs.mit.edu (Nate Osgood)>
250-<dean@anubis.network.com (Dean C. Gahlon)>
250-<fubar@sequent.com (Jay Vosburgh)>
250-<rjb@max.u.washington.edu (Richard Brzustowicz Jr)>
250-<voynich-incoming@castrov.cuc.ab.ca (Brett Wuth)>
250-<jkozak@cix.compulink.co.uk (John Kozak)>
250-<mathews@kong.gsfc.nasa.gov (Jason Mathews - 514)>
250-<cowan@snark.thyrsus.com (John Cowan)>
250-<Michael.Roe@cl.cam.ac.uk (Mike Roe)>
250-<acase@reed.edu (Andrew Case)>
250-<matt@kanatek.ocunix.on.ca (Matthew Harding)>
250-<dtl8v@holmes.acc.Virginia.EDU (Heraclitus)>
250-<foxd@silver.ucs.indiana.edu (daniel fox)>
250-<kornai@csli.stanford.edu (Andras Kornai)>
250-<mmonk@gmuvax2.gmu.edu (Martin Monk)>
250-<floehr@cis.ohio-state.edu (eric floehr)>
250-<rcarter@nyx.cs.du.edu (Ron Carter)>
250-<crispen@foxy.boeing.com (Bob Crispen)>
250-<firth@sei.cmu.edu (Robert Firth)>
250-<jim@island.COM (Jim Bisso)>
250-<bilako@trearn.bitnet (Ahmet Koltuksuz)>
250-<gcole@manta.nosc.mil (Guy Cole)>
250-<Doug.Brightwell@Corp.Sun.COM (Doug Brightwell)>
250-<wet!naga@cca.ucsf.EDU (Peter Davidson)>
250-<clark@uni2a.unige.ch (Robin Clark)>
250 <hudu@well.sf.ca.us (Scott Marley)>

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Sat Dec 28 00:18 PST 1991
Received: by gauss; Sat Dec 28 03:29:34 EST 1991
Received: by inet.att.com; Sat Dec 28 03:29 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Sat, 28 Dec 91 00:19:25 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Sat, 28 Dec 91 00:19:19 PST
Date: Sat, 28 Dec 91 00:18 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Dee catalogues; Clulee
To: voynich@rand.org
Message-Id: <C3CCA2E409BF80874D@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR


Clulee, p 327; paper back edition (1988 or so) listsm the following

Roberts, R J and Andrew G Watson. _John Dee's Library Catalogue. The
   Bibliographical Society.  Forthcoming.

I don't have a copy of Whitaker's handy, so I can't check to see if
it's actually been published.  I will try to check on this on Monday
or Tuesday (I suspect the library's not open this weekend).

Despite the fact that Clulee's treatment is very nice indeed, I
don't find a mention of Halliwell's _Private Diary_ of John Dee in
his bibliography.  French mentions it, so I got the citation
clearly enough  --Camden Society Publications vol XIX (1842).
my impression, from a memory of having seen this long ago, is that
it contains some version of a catalogue of Dee's library --
but I'm sure the Roberts & Watson will be by far the most useful.

From gauss!math.mit.edu!jbaez Sat Dec 28 14:23:06 EST 1991
Received: by gauss; Sat Dec 28 14:49:38 EST 1991
Received: by inet.att.com; Sat Dec 28 14:49 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Sat, 28 Dec 91 11:24:58 -0800
Received: from nevanlinna (NEVANLINNA.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA28619; Sat, 28 Dec 91 14:19:30 EST
Received: by nevanlinna; Sat, 28 Dec 91 14:23:06 EST
Date: Sat, 28 Dec 91 14:23:06 EST
From: jbaez@math.mit.edu
Message-Id: <9112281923.AA07105@nevanlinna>
To: voynich@rand.org
Subject: Dee's library catalog
Status: OR


Peter Davidson writes:

(ii)  There was "a catalogue of Dee's library, prepared in 1583, just
before a mob at Mortlake destroyed many of his books ..."  Is this
catalogue extant?  If so, does it list anything which might be the VMS?

Yes, this catalog is extent and fairly widely available.  As Jim Reeds
suspected, it has been reprinted.  I thought I posted the bibliographical
info on it but I can't find it in my records... perhaps I just sent it
to one of the Jims.  Anyway, it's in Widener library at Harvard, to which
I have access onyl when my fiancee is in town and in the mood to find things
for me.

P.D. also writes:

(iii)  Dee visited Prague several times during 1584-88, apparently the
only period during which he could have passed the MS to the Emperor Rudolf.

There is a diary of Dee extant (also at Widener) which might clarify this.

I feel confident that if Dee's library catalog or diary contained evidence
linking him with the Voynich, then D'Imperio or other assiduous voynichologists,
who have gone over the Dee theory, would have noted this fact.

jb

From gauss!gauss.att.com!reeds Sat Dec 28 15:10:01 EST 1991
Received: by gauss; Sat Dec 28 15:30:41 EST 1991
Received: by inet.att.com; Sat Dec 28 15:30 EST 1991
Received: from research.att.com by rand.org; Sat, 28 Dec 91 12:10:14 -0800
Message-Id: <9112282010.AA26883@rand.org>
From: reeds@gauss.att.com
Date: Sat, 28 Dec 91 15:10:01 EST
To: voynich@rand.org
Status: OR

Maybe its time for somebody to update the bibliography file.  There
have been lots of refs in the last few weeks which should be put into
pub/jim/biblio.  I am not volunteering for this one.

Jim Reeds.

From gauss!math.mit.edu!jbaez Sat Dec 28 16:09:27 EST 1991
Received: by gauss; Sat Dec 28 16:49:40 EST 1991
Received: by inet.att.com; Sat Dec 28 16:49 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Sat, 28 Dec 91 13:11:15 -0800
Received: from nevanlinna (NEVANLINNA.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA28729; Sat, 28 Dec 91 16:05:47 EST
Received: by nevanlinna; Sat, 28 Dec 91 16:09:27 EST
Date: Sat, 28 Dec 91 16:09:27 EST
From: jbaez@math.mit.edu
Message-Id: <9112282109.AA07269@nevanlinna>
To: voynich@rand.org
Subject: bibliography
Status: OR

I'll update the bibliography file.

John Baez


From gauss!rand.org!jim%mycroft Sat Dec 28 13:46:12 PST 1991
Received: by gauss; Sat Dec 28 16:57:24 EST 1991
Received: by inet.att.com; Sat Dec 28 16:57 EST 1991
Received: from mycroft.rand.org by rand.org; Sat, 28 Dec 91 13:46:15 -0800
Received: from localhost by mycroft.rand.org; Sat, 28 Dec 91 13:46:13 PST
Message-Id: <9112282146.AA16039@mycroft.rand.org>
To: voynich@rand.org
Subject: Book report -- The Private Diary of Dr. John Dee
Date: Sat, 28 Dec 91 13:46:12 PST
From: Jim Gillogly <jim%mycroft@rand.org>
Status: OR

I've read the Diary a couple of times -- they have it at the UCLA Library.
There isn't much that I see as relevant to the Voynich; as has been noted,
his *more* private diary was kept in other places, and some of it has been
found and lives in the British Library as the Sloane Manuscripts.

Here are a few selected entries.  Most is in English, some is in Latin,
and some is in English written with Greek characters.

1583.  Sept. 21st, we went from Mortlake, and so the Lord Albert Lasky,
I, Mr. E. Kelly, our wives, my children and familie, we went toward our
two ships attending for us, seven or eight myle below Gravessende.

[and the next entry is for 1586].

[Here's one from Trebona; looks like Edward Kelly making gold, then off
to Prague (Rudolf's court).]

1586. Dec 19th, 19die (novi kalendarii) ad gratificandum Domino Eduardo
Garlando, et Francisco suo fratri, qui Edouardus nuncius mihi missus
erat ab Imperatore Moschoviae ut ad illum venirem, E.K. fecit prolelem
lapidis in proportione unius...... gravi arenae super quod vulgaris oz.
et 1/2 et producta est optimi auri oz. fere: quod aurum post distribuimus
a crucibolo una dedimus Edouardo.
Dec. 30th, E.K. versus Pragam.

1587.
Jan. 18th, rediit E.K. a Praga.  E.K. browght with him from the Lord
Rosenberg to my wyfe a chayne and juell estemed at 300 duckettes; 200
the juell stones, and 100 the gold.  Jan. 21st, E.K. again to Prage and so
Poland ward.  Feb. 5th, I tok a jornay of myself from Trebon to Newhowse,
two myles of, to mete my Lord to comen with him.  I toke two horsemen of the
cyty with me. [...] Feb 19th, E.K. cam from Poland abowt none to Trebone:
I sent word to my Lord straight. [...] March 7th, E.K. dedit nobis 300
ducata.  Recepimus a Domino Illustrissimo [Lasky -- JJG] 3300. [...]
March 21st, E.K. gave me 170 more, and of the 200 for changing 60 remayne.
Contumelie et contemptus a Cholek et a Schonberg.  March 23rd, venimus
Trebonam.

[skipping on down past a fair number of alchemical experiments]

1588.
May 10th, E.K. did open the great secret to me, God be thanked!

------

Anyway, there is a great deal of to-ing and fro-ing on the Continent, but
no smoking guns that I've been able to find.  Dee appears more interested
in the alchemy than in mysterious manuscripts.  However, as we see in
the dealings with angels, also taking place during this period but not
recorded in the Private Diary, he was also heavily into the mysteries...

Jim Gillogly

From gauss!math.mit.edu!jbaez Sat Dec 28 17:38:26 EST 1991
Received: by gauss; Sat Dec 28 18:09:38 EST 1991
Received: by inet.att.com; Sat Dec 28 18:09 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Sat, 28 Dec 91 14:40:15 -0800
Received: from nevanlinna (NEVANLINNA.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA28846; Sat, 28 Dec 91 17:34:48 EST
Received: by nevanlinna; Sat, 28 Dec 91 17:38:26 EST
Date: Sat, 28 Dec 91 17:38:26 EST
From: jbaez@math.mit.edu
Message-Id: <9112282238.AA07332@nevanlinna>
To: voynich@rand.org
Subject: facts, please?
Status: OR

If someone sends me information about

Clulee on Dee (even the title would be nice!!)

Dee's diaries

--- I mean bibliographical information --- I'll pop it in the bibliography
I'm compiling.

jb


(Also, anything else you think belongs in the bibliography that's not in
the present version!)

From gauss!math.mit.edu!jbaez Sat Dec 28 18:05:08 EST 1991
Received: by gauss; Sat Dec 28 18:17:25 EST 1991
Received: by inet.att.com; Sat Dec 28 18:17 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Sat, 28 Dec 91 15:06:56 -0800
Received: from nevanlinna (NEVANLINNA.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA28877; Sat, 28 Dec 91 18:01:29 EST
Received: by nevanlinna; Sat, 28 Dec 91 18:05:08 EST
Date: Sat, 28 Dec 91 18:05:08 EST
From: jbaez@math.mit.edu
Message-Id: <9112282305.AA07341@nevanlinna>
To: voynich@rand.org
Subject: too late...
Status: OR

I have compiled my bibliography, incomplete as it is, and any information
you give me I will save for another updated version.  I will send my
bibliography to Jim Gillogly for it to be added to the files at rand.org.

jb

From gauss!cca.ucsf.EDU!wet!naga Sun Dec 29 20:12 PST 1991
Received: by gauss; Mon Dec 30 06:49:55 EST 1991
Received: by inet.att.com; Mon Dec 30 06:49 EST 1991
Received: from cgl.ucsf.EDU by rand.org; Mon, 30 Dec 91 03:32:29 -0800
Received: from cca.ucsf.EDU by cgl.ucsf.EDU (5.65/GSC4.21)
	id AA00482 for voynich@rand.org; Mon, 30 Dec 91 03:32:22 -0800
Received: from wet.UUCP by cca.ucsf.EDU (5.61/GSC4.21)
	id AA00316; Mon, 30 Dec 91 03:31:44 -0800
Received: by wet.uucp (/\=-/\ Smail3.1.18.1 #18.2)
	id <m0kyEMD-0002RXC@wet.uucp>; Sun, 29 Dec 91 20:12 PST
Message-Id: <m0kyEMD-0002RXC@wet.uucp>
Date: Sun, 29 Dec 91 20:12 PST
From: wet!naga@cca.ucsf.EDU (Peter Davidson)
To: voynich@rand.org
Subject: Further queries on Dee chronology
Status: OR

 
Thanks to Jacques Guy for his interesting review of Donald Laycock's
"The Complete Enochian Dictionary".  The following hypothesis is
certainly an intriguing possibility:
 
>                   ... pers. com. from Donald Laycock: there is,
>in Dee's diaries, a note to this effect: "How strange, I cannot
>find my Book of Soyga". According to Don, Dee's "Book of Soyga" may
>have been the Voynich manuscript, and Kelley would have stolen it and
>sold it to Rudolph II, emperor of Bohemia.
 
Is there a date available for this diary entry?
 
I'd like to get clearer on some chronology.  Jacques gives Kelley's
dates as 1555-1595 and writes:
 
>                       ...  Dee and Kelley spent much of their later
>years in eastern Europe (Poland and what is now Czechoslovakia).
>Kelley died in Bohemia under obscure circumstances, whereas Dee went
>back to England.
 
According to Brumbaugh, Dee was in Czechoslovakia only during the
years 1584-1588.  Do other sources confirm this?  Does this mean that
Kelley visited Czechoslovakia without Dee during the years 1588-1595?
 
I've heard a story that both Dee and Kelley were confined by the Emperor
Rudolf with firm instructions to make gold (which they were not able to do)
and that they escaped, Kelley dying in the attempt.  But if Dee left
Czechoslovakia for the last time in 1588 and Kelley died in 1595 then this
story cannot be correct.
 

From gauss!cca.ucsf.EDU!wet!naga Mon Dec 30 05:33 PST 1991
Received: by gauss; Mon Dec 30 23:30:32 EST 1991
Received: by inet.att.com; Mon Dec 30 23:30 EST 1991
Received: from cgl.ucsf.EDU by rand.org; Mon, 30 Dec 91 20:13:51 -0800
Received: from cca.ucsf.EDU by cgl.ucsf.EDU (5.65/GSC4.21)
	id AA16102 for voynich@rand.org; Mon, 30 Dec 91 20:13:46 -0800
Received: from wet.UUCP by cca.ucsf.EDU (5.61/GSC4.21)
	id AA07695; Mon, 30 Dec 91 20:13:04 -0800
Received: by wet.uucp (/\=-/\ Smail3.1.18.1 #18.2)
	id <m0kyN7G-0003G2C@wet.uucp>; Mon, 30 Dec 91 05:33 PST
Message-Id: <m0kyN7G-0003G2C@wet.uucp>
Date: Mon, 30 Dec 91 05:33 PST
From: wet!naga@cca.ucsf.EDU (Peter Davidson)
To: voynich@rand.org
Subject: Foiled again
Status: OR

 
GLADIS, U.C. Berkeley, says:
 
    Your search for the Personal Name: LAYCOCK DONALD C
    retrieved 2 records.
 
    1.     Laycock, Donald C.
           The complete Enochian dictionary : a dictionary of the Angelic
             language as revealed to Dr. John Dee and Edward Kelley / by
           London : Askin Publishers Ltd. ; New York : distributed by S.
             Weiser, 1978
 
    Main Stack   PM9021.E55.L39
                   MISSING
 
Looks like the conspiracy got there first.
 

From gauss!trl.OZ.AU!j.guy Mon Dec 30 08:51:19 EST 1991
Received: by gauss; Sun Dec 29 18:49:56 EST 1991
Received: by inet.att.com; Sun Dec 29 18:49 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Sun, 29 Dec 91 13:51:36 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA00380; Mon, 30 Dec 91 08:51:19 EST
Date: Mon, 30 Dec 91 08:51:19 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112292151.AA00380@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Laycock's Enochian Dictionary
Status: OR


                             A review of

                   The Complete Enochian Dictionary
                          by Donald Laycock
                   (London: Askin Publishers, 1978)


John Baez's call for a review of Laycock's "Enochian Dictionary"
prompts me to write this. As I wrote in one of my postings, I had
known Donald Laycock for some fifteen years, the duration of my stay
with the Linguistics Department of the Australian National University
in Canberra, where I first came as a Ph.D. student in 1970 and stayed
on until 1985. Laycock, who was born in 1936 and died of leukemia in
1986, was a specialist in the languages of Papua-New Guinea, which
line he considered as an all too often boring meal ticket. "What I'm
interested in" he used to say, "is writing poetry and inventing
puzzles". He was also skeptically interested in things occult and
strange and it is from him that I learnt of the Voynich manuscript. I
used to marvel at how he would spend considerable time and effort
analyzing stuff that he himself would brand as nonsense. Neither first
nor foremost, Dee's "Angelic" language, Enochian.

Laycock's "Complete Enochian Dictionary" is a 270-page hard-bound
volume. It is based on several months of poring over dusty volumes in
grand old dusty reading rooms of the British Museum. That was Don's
idea of pleasantly spending part of a sabbatical -- in 1975 or
thereabouts if memory serves me. What is Enochian? It is the language
of angels, as revealed to Dr. John Dee (1527-1608), scholar and
astrologer to the court of Queen Elizabeth I, through a medium, Edward
Kelley (1555-1595).

Pages 7 to 18 of Laycock's Enochian Dictionary are taken up by a
preface by Stephen Skinner, dated 1975, three years before the
publication of the Dictionary.

The next 46 pages are a terse summary of Don's sleuthing foray into
Dee's diaries, sensible and less sensible sources for the Enochian
language, whether it could be a cipher (spoiler: it cannot possibly
be), and most importantly in my view, a phonetic and grammatical
analysis of the Enochian language, showing that angels (for Enochian
is purported to be the language of angels and of Adam before the fall)
used a language the spelling and syntax of which were strikingly
similar to Elizabethan English.

The rest is a two-way dictionary, Angelic-English and English-Angelic,
followed by a complete corpus of all the Enochian texts transmitted to
Dee via Kelley, with interlinear translations, based on Dee's notes.
Finally, three pages of bibliography: every source you could ever find
concerning Enochian, Dee, and Kelley.

In building his dictionary Don did not only take into account Dee's
Enochian corpus, but the much later versions used in the rituals of
the Golden Dawn and Aleister Crowley's. Entries are cross-referenced
where appropriate, thus for instance this entry:

kanila       AC:G = cnila, blood.
  canilu     AC:C

Turning to entry "cnila", we read:

cnila     kni-la  blood.
  kanila  AC:G
  canilu  AC:C

where "kni-la" gives the pronunciation, as reconstructed from Dee's
diaries, and kanila and canilu the forms found as "Variants of
Enochian words in the goetic invocations published by Aleister Crowley
as a supplement to MacGregor Mathers' translation of the Goetia"
(AC:G) and in "Aleister Crowley's version of the Calls, with
additional magical data, as published in The Equinox, Vol.I, Nos. 7 &
8" (AG:C). Don even went to the trouble of comparing the Enochian of
La Vey's "Satanic Bible" (which I had ferretted out of the shelves of
Dalton's bookshop in beautiful downtown Canberra):

"Another source of Enochian vocabulary, La Vey's Satanic Bible, was
examined, but the Enochian here follows AC:C exactly, with only two
additional misprints (mamao for momao, and gianai for ginai) which
were thought too insignificant to be included in the Dictionary").

The dictionary part of "The Complete Enochian Dictionary", then, is as
serious and scholarly a piece of work as could ever be produced,
never mind that its object was, in Don Laycock's very own words, a
fabrication by a rogue and a blackguard (Edward Kelley).

Yet, to me, upon reading this book which I had never carefully read
before, all its value lies in the fifty pages or so where Don
introduces John Dee and Edward Kelley, and tells the story of
Enochian. In there you will find the most amazing treasure trove of
information, from the very earliest form of angelic communication, on
March 20, 1582, to an early experiment in wife swapping
(angel-sponsored) via a prophecy of the demise of the Spanish Armada
and of the execution of Mary Queen of Scots, and how Kelley asked a
young girl spirit if she could lend him a hundred pounds for a
fortnight (I kid you not!).

What, will you ask, has all that to do with the Voynich manuscript? At
the most, little I dare say. Dee and Kelley spent much of their later
years in eastern Europe (Poland and what is now Czechoslovakia).
Kelley died in Bohemia under obscure circumstances, whereas Dee went
back to England. On p.170 of the Complete Enochian Dictonary, there
is this entry:


soyga         will of God (?) (The Book of
              Soyga was one of Dee's holy
              books; the spirits said it was
              not a reversal of the Greek
              agios, holy.)

to which you must add this pers. com. from Donald Laycock: there is,
in Dee's diaries, a note to this effect: "How strange, I cannot
find my Book of Soyga". According to Don, Dee's "Book of Soyga" may
have been the Voynich manuscript, and Kelley would have stolen it and
sold it to Rudolph II, emperor of Bohemia.


From gauss!gauss.att.com!reeds Mon Dec 30 11:20:03 EST 1991
Received: by gauss; Mon Dec 30 14:55:02 EST 1991
Received: by inet.att.com; Mon Dec 30 14:55 EST 1991
Received: from research.att.com by rand.org; Mon, 30 Dec 91 11:27:45 -0800
Message-Id: <9112301927.AA16859@rand.org>
From: reeds@gauss.att.com
Date: Mon, 30 Dec 91 11:20:03 EST
To: voynich@rand.org
Status: OR

Here is a little essay.   Comments?

The Phonetic Fallacy
Jim Reeds
30 Dec 1991

1.  Encipher a paragraph or so of any language normally written in Roman
letters, with a simple substitution where vowels stand for vowels and
consonants stand for consonants.  The resulting cipher text looks vaguely
like plain text in an unknown (but pronounceable) language.

For example, using the following mapping

	plain:	AEIOU BCDFGHJKLMNPQRSTVWXYZ 
	cipher: UIOEA TVWXYZBCDFGHJKLMNPQRS

this plain text

   exstant et ad ciceronem, item ad familiares domesticis de rebus, in
   quibus, si qua occultius perterenda erant, per notas scripsit, id est
   sic structo litterarum ordine, ut nullam verbum effici posset; quae si
   qui investigare et persequi velit, quartem elementorum litteram, id est
   d pro a et perinde reliquas commutet.

   there are also letters of his  to cicero, as well as
   to his intimates on private affairs, and in the latter, if he had
   anything confidential to say, he wrote it in cipher, that is, by so
   changing the order of the letters of the alphabet, that not a word
   could be made out.  if anyone wishes to decipher these, and get at
   their meaning, he must substitute the fourth letter of the alphabet,
   namely d, for a, and so with the others.

yields this cipher text

   iqlmugm im uw vovikegif, omif uw xufodoukil wefilmovol wi kital, og
   jaotal, lo jau evvadmoal hikmikigwu ikugm, hik gemul lvkohlom, ow ilm
   lov lmkavme dommikukaf ekwogi, am gadduf niktaf ixxovo hellim; jaui lo
   jao ognilmoyuki im hiklijao nidom, jaukmif idifigmekaf dommikuf, ow ilm
   w hke u im hikogwi kidojaul veffamim.

   mziki uki udle dimmikl ex zol  me vovike, ul pidd ul
   me zol ogmofumil eg hkonumi uxxuokl, ugw og mzi dummik, ox zi zuw
   ugrmzogy vegxowigmoud me lur, zi pkemi om og vohzik, mzum ol, tr le
   vzugyogy mzi ekwik ex mzi dimmikl ex mzi udhzutim, mzum gem u pekw
   veadw ti fuwi eam.  ox ugregi polzil me wivohzik mzili, ugw yim um
   mziok fiugogy, zi falm latlmomami mzi xeakmz dimmik ex mzi udhzutim,
   gufidr w, xek u, ugw le pomz mzi emzikl.

An ignorant person might suppose the encrypted Latin was really a sample
of a heretofore unknown Germanic language, and make tentative identifications:

	``dommikuf'' = cognitively challenged person
	``dommikukaf' = stupid cake
	``jaotal'' =  the name of the language 
	``og'' = and

The word ``Jaotal'' could be explained:  Jao is the Portugese for our
John, or German Hans.  ``Jaotal'' literally means ``John's, or Hans's,
tounge'', that is, the speech of the Hanseatic league.  And so on.

On the same sort of evidence, the second paragraph is Swahili, with 
English borrowings, like ``latlmomami'' = ``my little mother'', or ``I am
a little mother''.



2. The foregoing is obviously a reflection of the fact that most of (all of?)
the languages we know exhibit some gross phonological similarites, among which
is the approximate alternation of vowels and consonants.  Any string of letters
with some degree of vowel/consonant alternation is pronounceable, and so might
be supposed to be the written form of an utterance in some unknown language.


3. This is a pitfall for unwary would-be Voynich manuscript decipherers.

The text of the Voynich MS does display a patterning of letters like
vowel/consonant alternation.  Following Jacques Guy, we may recover
putative vowels by examining digraphic frequency tables.  The letters O9ASC
in Currier's transliteration seem vowel-like, that is, are very 
frequently found adjacent to ``consonants''.  If we apply the following
haphazardly-chosen simple substitution

	cipher (Currier):	OA9SC Z84MDFRQPE2J
	plain:			oaeui lrhtpdgvxfny

to some arbitrarily chosen A and B text from the VMS, we obtain

	Vauen edaf ag axat lof loge vogen e dog lofre 
	noge Xag o g e daN lxat lag ag* vag vag rap 
	neaU lide og edat lor voage v*g ragat ne 
	$ot oxiie oxiog nofoxe v*ag rat odat of odap 
	naT e uiag vat Wag Yat 
	eragaIlo 
	* orag *e lof Woe oerag l* n Yoat lorane 
	*le uore oduoe oxuof uove on ue raN log don 
	rat lon Yof lore 
	raN *oxiore 

and

	Bliode oraU hoe oVlor ueBuie eBuireaN uoVo uiBure 
	ruie nat a7iiore edivie uire exire re uivire fn 
	oat lve vire ofoe edlre ofuire ref eliie nat n 
	hodire uion exire hodire exire uidire rat oray nafre 
	nat lire iire iire nue rat vire hodiire hodire vie 
	guire hodire 
	Biio hodire rag lio eBunirn n at laBuire Vue raf uire nag 
	rat lire hodiire hoxiirag n odof exire hodiire h3dire 
	xuiole rlre odire uXe n re re ediiue odiire uide 
	lini at lion uiore oxaf 

which is prounouncible.  (Some Currier letters -- WBVN73 --  are not mapped.)

The substitution used here was chosen only to make Roman vowels stand
for putative Voynich vowels.  If one wanted, the particular assignment of
Roman values to Voynich letters could be ``fine tuned'' to make almost
any theory about the ``plain text'' language of the VMS seem more plausible,
especially when only a small portion of the corpus is examined and when
the decipherer allows himself the latitude of inventing new dialects
(as does Levitov, of Flemish),  or of declaring the plain text to be quaintly 
misspelled (Feeley), or also largely gibberish anyway (as does Brumbaugh).

The temptation to discover that the VMS is written in Jaotal has led more
than one researcher astray.


From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Mon Dec 30 16:44 PST 1991
Received: by gauss; Mon Dec 30 20:19:12 EST 1991
Received: by inet.att.com; Mon Dec 30 20:19 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Mon, 30 Dec 91 16:45:56 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Mon, 30 Dec 91 16:45:49 PST
Date: Mon, 30 Dec 91 16:44 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: the place of play in a world of work
To: voynich@rand.org
Message-Id: <C1B077EC95BFA024EF@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

In a speculative message, J Reeds asks for comments.
     From:	IN%"reeds@gauss.att.com" 30-DEC-1991 
     12:01
     To:	voynich@rand.org
     Message-id: <9112301927.AA16859@rand.org>
     X-Envelope-to: rjb
     Here is a little essay.   Comments?
     The Phonetic Fallacy
     Jim Reeds
     30 Dec 1991
     1.  Encipher a paragraph or so of any language 
     normally written in Roman letters, with a simple 
     substitution where vowels stand for vowels and 
     consonants stand for consonants.  The resulting 
     cipher text looks vaguely like plain text in an 
     unknown (but pronounceable) language.
             [ example deleted ]


     2. The foregoing is obviously a reflection of the 
     fact that most of (all of?) the languages we know 
     exhibit some gross phonological similarites, among 
     which is the approximate alternation of vowels and 
     consonants.  Any string of letters with some 
     degree of vowel/consonant alternation is 
     pronounceable, and so might be supposed to be the 
     written form of an utterance in some unknown 
     language.

Well, Polish, or for that matter romanized Tibetan, 
would look very different from (say) Hawaiian.  It's 
quite possible that a Polish cryptologist applying this 
maneuver to a Hawaiian encrypted text might find the 
vowel/consonant ratio quite bewildering.  Or vice 
versa.  In other words, in addition to pronunciation 
(with its approximate alternation of vowels and 
consonants) there is *spelling*, and spelling 
conventions can be quite odd.  Imagine the English word 
"knight," spelled perhaps "nayit" and then encrypted.  
Or "Huckleberry Finn," or for that matter "Hermann 
Hesse," or "You shouldn't take your pleasure in the 
church at noon," spelled as a Russian would in 
Cyrillic, then romanized and encrypted.  Or, more 
plausibly for the Voynich, French spelled phonetically 
and then encrypted.

     3. This is a pitfall for unwary would-be Voynich 
     manuscript decipherers.
     The text of the Voynich MS does display a 
     patterning of letters like vowel/consonant 
     alternation.  Following Jacques Guy, we may 
     recover
     putative vowels by examining digraphic frequency 
     tables.  The letters O9ASC in Currier's 
     transliteration seem vowel-like, that is, are very 
     frequently found adjacent to ``consonants''.  If 
     we apply the following haphazardly-chosen simple 
     substitution ... chosen only to make Roman vowels 
     stand for putative Voynich vowels.  If one wanted, 
     the particular assignment ofRoman values to Voynich 
     letters could be ``fine tuned'' to make almost
     any theory about the ``plain text'' language of 
     the VMS seem more plausible, especially when only a 
     small portion of the corpus is examined
     and when the decipherer allows himself the 
     latitude of inventing new dialects (as does Levitov, 
     of Flemish),  or of declaring the plain text to be
     quaintly misspelled (Feeley), or also largely 
     gibberish anyway (as does Brumbaugh).The temptation 
     to discover that the VMS is written 
     in Jaotal has led more than one researcher astray.

Exactly -- which is why this sort of attempt to resolve 
the cipher (if there is one) should never be taken 
seriously.  This doesn't mean they should be avoided.
They should be taken as play:  diversions that just 
might reveal something interesting, but which are really 
justified by their ability to entertain.

From gauss!gauss.att.com!reeds Mon Dec 30 19:49:24 EST 1991
Received: by gauss; Mon Dec 30 20:13:43 EST 1991
Received: by inet.att.com; Mon Dec 30 20:13 EST 1991
Received: from research.att.com by rand.org; Mon, 30 Dec 91 16:49:31 -0800
Message-Id: <9112310049.AA21613@rand.org>
From: reeds@gauss.att.com
Date: Mon, 30 Dec 91 19:49:24 EST
To: voynich@rand.org
Status: OR

The current issue of Cryptologia (October 1991, vol XV, No 4) mentions
(then) Lt. Prescott Currier on p.287 as greatly helping Agnes Driscoll
break JN25 in 1939/40.

From gauss!Csli.Stanford.EDU!kornai Mon Dec 30 19:51:25 PST 1991
Received: by gauss; Mon Dec 30 23:37:52 EST 1991
Received: by inet.att.com; Mon Dec 30 23:37 EST 1991
Received: from Csli.Stanford.EDU by rand.org; Mon, 30 Dec 91 19:51:32 -0800
Received: by Csli.Stanford.EDU (4.1/inc-1.0)
	id AA16973; Mon, 30 Dec 91 19:51:25 PST
Date: Mon, 30 Dec 91 19:51:25 PST
From: Andras Kornai <kornai@Csli.Stanford.EDU>
To: voynich@rand.org
Subject: Phonetic non-fallacy
Message-Id: <CMM.0.90.2.694151485.kornai@Csli.Stanford.EDU>
Status: OR

Let me comment on Jim Reeds' little essay on the "Phonetic fallacy"
from a somewhat different perspective. Having received a couple of
pages from the ms from Jim Gillogly a few days ago (Thanks, Jim!), I
was struck by two characteristics of the text: first, how little there
is (on the basis of these two pages I estimate the whole document to
be less than 200K characters) and second, how repetitive it is.

This repetitiveness is so glaring that any reasonable theory must 
take it as its starting point. One possibility that leaps to mind is 
that the text is not prose but verse, of a highly repetitive, if
you wish, mantric character. Would it be unusual for a manuscript 
like this to contain not only the pictures of plants etc. but also 
terse instructions (e.g. how to extract medicines from them) and
also magic spells to be chanted during this process (or to enhance
the power of the medicine, etc.)? Would it be unheard of?

There is another explanation of the repetitiveness which runs as
follows. In India there is a well-known tradition of committing large
amounts of text to memory: whole dramas, epics, and scientific
treatises are transmitted that way from generation to generation.
Repetition is used as an important memory aid in this process: to
remember ABCDEFG the student will chant ABCBCDCDEDEFEFG, or
ABCBABCDCB... or some similar repetitive variatiant of the original
text. (The results are even better for Sanskrit, because the sandhi
will introduce slight alterations, so that the B in ABC will often
come out a little differently from the B in CBC).

I'm NOT claiming that the ms contains mantras, spells, or other
verse material, nor do I claim that it contains repetitivized
text: however, I reagard both as plausible hypotheses that can,
indeed must, be tested before we can ascertain the language (if
any) of the manuscript. Therefore, I find it imperative that 
we settle on a transcription system that (unlike Currier or the
others) is pronuncable (i.e. the resulting text can be read aloud).
What Jim presented as a "phonetic phallacy" is in fact a good 
first start: it is *much* better to read

>	Vauen edaf ag axat lof loge vogen e dog lofre 
>	noge Xag o g e daN lxat lag ag* vag vag rap 
>	neaU lide og edat lor voage v*g ragat ne 
>	$ot oxiie oxiog nofoxe v*ag rat odat of odap 
>	naT e uiag vat Wag Yat 
>	eragaIlo 
>	* orag *e lof Woe oerag l* n Yoat lorane 
>	*le uore oduoe oxuof uove on ue raN log don 
>	rat lon Yof lore 
>	raN *oxiore 

than anything else I have seen so far. Why is it important to
read the text, even at the price of possibly mispronuncing it?
Because the human mind is a better pattern recognizer than anything
else we got (FAR better than computers, as we all know), and this
whole enterprise is about discerning a pattern. To find the rhythmic 
pattern (which must be present if either of the above hypotheses 
hold) we have to be able to read the stuff aloud, so Currier won't 
do.

Needless to say, the enterprise of making the text pronuncible is not
to be confused with the larger enterprise of making sense of it.  Here
there are obvious adequacy criteria, such that essentially the whole
corpus must be rendered pronuncible, not just select passages, all
letters need be mapped, there can be only a handful of typos, etc.
Hypotheses can be formed on the basis of the material now available 
and tested as more pages become available (to those of us who still 
don't have the entire text): are these new pages also pronuncible?

But on the whole, I find Jim's first attempt *very* encouraging.
First of all, it is almost completely successful as is, and assumes
the typologically most frequent vowel system (the 5 vowels aiueo, just
as in Latin). Other vowel systems, such as 7 or 8 vowels are also
quite common (for a good survey, see Ian Maddieson's "Patterns of
Sounds", Cambridge University Press 1984) but if we can get 5 to
work that would be the best. Second of all at the moment it is 
easy to tweak so that minor blemishes (such as the freestanding g 
in line 2) are removed: for instance we can just use "s" instead of 
"g". 

In general, there are a number of subtle asymmetries in the segment
inventories of the languages of the world that will act to remove 
a great deal of the indeterminacy of the substitution: for instance
stops (such as p,t,k) generally do not appear in isolation
while vowels, glides, liquids often do. At any rate, the fact that
a first attempt can come so close is extremely encouraging: I feel
virtually certain that we do not have a mora-, syllable-, or morpheme-
based writing system but a vanilla phoneme-based alphabet. (The
text is surprisingly vanilla in other respects as well, such as 
obviously left-to-right top-to-bottom writing.)

Suppose, for example, that the whole ms can be pronunced under a 
certain substitution S, and not only can it be pronunced, but 
in fact yields perfect hexameters when read aloud. We still can 
not understand it, but if all substitutions S' that yield good
meter assume 8 vowels, we have good evidence that the text was
written in Turkish or some other language with an 8-vowel system.
If on the other hand a 5-vowel system works well, the text can
not be Turkish but can be Latin. I'll start work on this as soon
as the maintainer of the archive will de me the favor of randomly
cutting the corpus (both "hands") in two parts: I will first 
ftp one part (the training data) and make a hypothesis, and only
then will I fetch the second part (the test data) to see if my
substitution extends to it.

Andras Kornai



From gauss!cca.ucsf.EDU!wet!naga Tue Dec 31 01:14 PST 1991
Received: by gauss; Tue Dec 31 11:24:51 EST 1991
Received: by inet.att.com; Tue Dec 31 11:24 EST 1991
Received: from cgl.ucsf.EDU by rand.org; Tue, 31 Dec 91 03:40:31 -0800
Received: from cca.ucsf.EDU by cgl.ucsf.EDU (5.65/GSC4.21)
	id AA20886 for voynich@rand.org; Tue, 31 Dec 91 03:40:23 -0800
Received: from wet.UUCP by cca.ucsf.EDU (5.61/GSC4.21)
	id AA11096; Tue, 31 Dec 91 03:39:37 -0800
Received: by wet.uucp (/\=-/\ Smail3.1.18.1 #18.2)
	id <m0kyfYS-0000UhC@wet.uucp>; Tue, 31 Dec 91 01:14 PST
Message-Id: <m0kyfYS-0000UhC@wet.uucp>
Date: Tue, 31 Dec 91 01:14 PST
From: wet!naga@cca.ucsf.EDU (Peter Davidson)
To: voynich@rand.org
Subject: Meaning/gibberish
Status: OR

 
Some thoughts stimulated by Jim Reeds' essay:
 
Scanning the following text (which consists of ten sentences) will
immediately reveal regularities.  A detailed analysis will reveal a
limited vocabulary and an equally limited syntactical structure.  Without
too much work (given further samples of text, and I can supply as much
as needed) the entire "language" could be specified exactly, as regards
vocabulary and syntax.
 
 Tle soalifunt's mintrogger asta findagleg dringmen kerpiz qtoolte fogler
 lenmyzs et findagleg ranmfos trilknowia rpook algmikox sed withi tolunet
 alofgs nofeif yueep afth incominatlef dringmen. Harft galdranent firkels
 malxow tiko qtoolte menkolacis sofxiers et findagleg ranmfos samt fo u
 ksile trang tanu semik ano fo ghriakll sed ule senlocs kerpiz rukalfte
 fogler sofxiers!  Fid deu ahof findagleg Sodas ren key samt fo u ksile nom
 tu menkolacis galdranent firkels et harft tolunet alofgs ollcstiokapi siif
 takir seni tanokon sed findagleg tolunet alofgs kerpiz bauqokievt
 findagleg crontis - eferith tu fogan tle mukjilt tu ab kikge. I negi asta
 findagleg gekbrals fenally yueep afth ule lenmyzs et conpreft senlocs
 kerpiz stholt ighthymuiron sed menkolacis Sodas ren key ollcstiokapi
 bauqokievt ule anbors - eferith tu fogan qikody mifen akboksi wuihol
 srenkifemt uokklos. Kn's reffit akanobege dhat harft lenmyzs samt fo u
 ksile yueep afth incominatlef Sodas ren key et conpreft senlocs samt fo u
 ksile menaki ighthymuiron sed conpreft galdranent firkels trilknowia
 axfote ule sofxiers - eferith tu fogan qikody mifen akboksi wuihol
 srenkifemt uokklos. Co dan mif, harft dringmen fenally qtoolte spofdrintig
 Sodas ren key et ule galdranent firkels ollcstiokapi yuokt tanu semik ano
 fo ghriakll sed withi tolunet alofgs samt fo u ksile uyamer im spofdrintig
 tolunet alofgs! Tle murga nix fogler gekbrals nofeif nom tu menkolacis
 dringmen et conpreft dringmen kerpiz alfi xille kprooft sed spofdrintig
 Sodas ren key fenally axfote withi lenmyzs - "Bob" bilo ax. Menkolacis
 galdranent firkels samt fo u ksile bauqokievt withi galdranent firkels et
 incominatlef tolunet alofgs samt fo u ksile oog urkthmaas sed harft
 sofxiers toraki rukalfte withi sofxiers - tou fan yuke igo nzid yto ik!
 Fid deu ahof harft crontis malxow tiko bauqokievt ule crontis et
 incominatlef senlocs fenally inace xille kprooft sed spofdrintig senlocs
 toraki truse incominatlef sofxiers - eferith tu fogan tle mukjilt tu ab
 kikge. Co dan mif, ule ranmfos asmia nom tu menkolacis gekbrals et withi
 sofxiers samt fo u ksile tokt urkthmaas sed ule stoll broxyzs asmia chiki
 fogler crontis.
 
The syntactical structure can be revealed exactly, and a persuasive case
could be made that the language mostly consists of adjectives, verbs,
nouns and connectives very similar to English.  In fact the passage was
generated by a computer program that I wrote, and the adjectives, etc.,
are quasi-random combinations of letters, and have no meaning.  Since
they have no meaning the text as a whole is meaningless.  Unless, of
course, meaning is given to the words.  For example, "Sofxiers" might be
taken to have a common origin with "soldiers", and "gekbrals" might be
taken to mean "generals".  If each term is associated with an English
term in the corresponding grammatical class then grammatically correct
and comprehensible (if not rational) English can be produced.  If "good
reasons" can be given for translating the terms in that particular way
then we have a "decipherment" of the gibberish above.
 
Although this text was generated by a computer program, the same kind of
thing could be generated by a human brain in an appropriate (if somewhat
unusual) condition.
 
Could the Voynich ms. be the same kind of artifact?  How could we disprove
the hypothesis that it is?  Only, it seems to me, by identifying a set of
rules which could be used to convert the text into text within a known
language (English, Latin, etc).  This is, of course, what "decipherment"
means.  The absence so far of the discovery of such a set of rules is
evidence in favor of the hypothesis that the Voynich ms. is meaningless.
On the other hand, it could be argued that there just hasn't been enough
effort (of the right kind) expended.  Have the world's best linguists
(besides Jacques) studied the text seeking similarities to known lexical
patterns in Romansch, Catalan, Gaelic, Latvian, Georgian, Anatolian,
Arapaho, Blackfoot, etc.?  If not, and in the absence of a successful
connection to some language, can we ever rule out the possibility that
some linguist may some day discover by statistical analysis compelling
similarities between the Voynich text and some obscure language which
would permit a translation?  Only, it seems, on extra-linguistic
evidence, such as an entry in Dee's diary along the lines of: "Today
completed with E.K. and the blessings of the divine spirits the
all-too-laborious concoction of the meaningless garbage that we shall
pass off tomorrow to H.R.E. for 6000 ducats."
 
So my question for further consideration is:  If Dee, Kelley or some
other person deliberately concocted a text which was meaningless
garbage, could this ever be proved on the basis of internal evidence?
 
If they did, and if not, and if there is no decisive external evidence,
then presumably the mystery of the Voynich ms. will remain forever unsolved.
 
 
P.S.  Synchronistically, as I logged on to upload this I was greeted by:
 
                    To A Quick Young Fox:
            Why jog exquisite bulk, fond crazy vamp,
            Daft buxom jonquil, zephyr's gawky vice?
            Guy fed by work, quiz Jove's xanthic lamp --
            Zow!  Qualms by deja vu gyp fox-kin thrice.
                            -- Lazy Dog
 

From gauss!trl.OZ.AU!j.guy Tue Dec 31 09:15:35 EST 1991
Received: by gauss; Mon Dec 30 17:33:30 EST 1991
Received: by inet.att.com; Mon Dec 30 17:33 EST 1991
Received: from medici.trl.OZ.AU by rand.org; Mon, 30 Dec 91 14:16:47 -0800
Received: from hal.trl.OZ.AU by medici.trl.OZ.AU (4.1/AIS-1M)
	id AA00367; Tue, 31 Dec 91 09:15:35 EST
Date: Tue, 31 Dec 91 09:15:35 EST
From: j.guy@trl.OZ.AU (Jacques Guy)
Message-Id: <9112302215.AA00367@medici.trl.OZ.AU>
To: voynich@rand.org
Subject: Jim Reed's Phonetic Fallacy
Status: OR





Very good! Very funny, too. The second Voynich interpretation looks
like Pidgin Italian to me. Verily, verily, thou hast the mystery
solved.  The Voynich manuscript was written in a Sabir dialect (Sabir
was an Italian-based trading jargon long used in the Mediterranean),
and in a simple substitution cipher.

Viz:

Biio hodire rag lio e Bunirn n at laBuire Vue raf uire nag...

Read: Vio  adire  Raglio e Burnini ad la Buire, vue rapire (nag?)

Vio (Spanish): he saw
adire: arrive
Raglio, Burnini: persons' names
la Buire: unidentified place name
vue: Spanish "fue" he was, he went.
rapire: to steal, kidnap (raf uire: a mixture of Italian "rapire" and French "ravir")

Oh, I know the rest: nag rat lire hodiire hoxiirag n odof exire

The word breaks are just wrong. Here they are, corrected:

nagr at lire, ho dire, ho XII ragn o devo exire

nagr is "negro"
at lire: is "for liras", that is, for money, for profit

ho: Italian "I have"
dire: Italian "to say"
ragn: is back-slang for nagr "negro" again
o: Italian "or"
devo: Italian "I must", or "I owe"
exire: to exit.

It's all very clear now:

He (I perhaps I) saw Raglio and Burnini arrive at la Buire. He/I had gone
to kidnap black slaves for money. I said: I have 12 slaves or I must quit.

"o" ("or") is probably a scribe's mistake for "e" ("and"). If so, the last
sentence becomes: I have twelve slaves (for sale) and I must leave (soon).

No wonder the Voynich is written in cipher! It is the memoires of
slavers who plying their trade somewhere in North Africa.  The pictures
are obviously red herrings. That explains the two handwritings, and
"languages", too: one is by the captain, the other by the first mate.
They spoke different dialects of Sabir, obviously.  Perhaps one was
Venetian and the other Maltese?



From gauss!gauss.att.com!reeds Tue Dec 31 11:17:45 EST 1991
Received: by gauss; Tue Dec 31 12:14:05 EST 1991
Received: by inet.att.com; Tue Dec 31 12:14 EST 1991
Received: from research.att.com by rand.org; Tue, 31 Dec 91 08:25:51 -0800
Message-Id: <9112311625.AA00207@rand.org>
From: reeds@gauss.att.com
Date: Tue, 31 Dec 91 11:17:45 EST
To: voynich@rand.org
Status: OR

It would be interesting to find the most pronouncible substitution 
from Currier to phonetic values, using, as Kornai suggests, information
about stops, glides, etc. I don't think we should transcribe in such a
``phonetic alphabet'' but it might well be worth our while to learn to
pronounce Currier letters phonetically.

I am sure that the VMS was uttered as it was written, whether or not
it is pure gibberish, gibberish with cipher text concealed in it,
or (but this is hard to believe) plain language.

Jim Reeds

From gauss!math.mit.edu!jbaez Tue Dec 31 12:54:18 EST 1991
Received: by gauss; Tue Dec 31 13:43:27 EST 1991
Received: by inet.att.com; Tue Dec 31 13:43 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Tue, 31 Dec 91 09:56:24 -0800
Received: from nevanlinna (NEVANLINNA.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA04263; Tue, 31 Dec 91 12:50:26 EST
Received: by nevanlinna; Tue, 31 Dec 91 12:54:18 EST
Date: Tue, 31 Dec 91 12:54:18 EST
From: jbaez@math.mit.edu
Message-Id: <9112311754.AA10663@nevanlinna>
To: voynich@rand.org
Subject: Dee, gibberish
Status: OR


Yesterday my friend Peter Durigan and I went to the rare book room at
Wellesley College (where I teach) to look at an old book by John Dee.
Peter is a patent lawyer with a very serious interest in Neoplatonism.
He once met an occult book dealer who said he had a copy of Dee's translation
of Euclid's "Elements," which he would sell for $15,000.  The book dealer
also said he would lend any book for inspection for 10 days, and Peter
proposed to me that we borrow Dee's "Elements," but I said it was a bad
idea, since 1) the book dealer might not really have meant that he'd lend
*that* book, and 2) what if we lost it?  So I looked to see what they had
of Dee's at Wellesley, and it turns out they have 1) the "Elementa" (as
it's really called), 2) "The Private Diary...," and 3) a catalog of Dee's
library (which text was listed as "under repair".   Clearly Wellesley was the
friend of a Dee fanatic at one time... perhaps someday I'll find out who.

Anyway, we arranged to see the "Elementa."  In part I was interested in the
text itself (since Peter and I have had many discussions of mysticism, 
Pythagoreanism, and Platonism as they relate to mathematics), and in part I 
wanted to learn the etiquette of rare book rooms on friendly ground before
tackling the Beinecke.  

The rare book room was on the fourth floor of the library.  We had to
ring a doorbell, and were buzzed in by the young woman I had spoken to
on the phone earlier.  We had to fill out forms indicating our names,
addresses, telephone numbers and affiliation with Wellesley (if any), and
I showed her my Wellesley I.D. and filled out a separate form saying which
book we were looking at.  She took the book over to a reading desk.  It was
a small leather-bound volume, published in the late 1500's by someone who
wrote in his prologue that Dee had died before being able to publish this
volume.  After this prologue there was a long preface by Dee, and then Dee's
translations of 6 books of the Elements into English.  Dee's preface
was fascinating.  There were illustrations on the first page showing
a naked woman (in a style utterly unlike those in the Voynich), and the
"D" of the first sentence, which began "Divine Plato..." was illuminated,
showing a figure of a lion rampant (with an erection, according to Peter),
a symbol which I interpreted as being the usual symbol for Mercury (though
it differed slightly from the standard version), and lots of abstract curlicues.
There was also a picture of two men, one firing an arrow...  It began with
a description of how many visitors came to Plato's academy and, finding that
nothing "practical" was taught therein, left in disgust, until Aristotle 
pointed this out and said that perhaps there should be some sort of 
notice warning them what it was all about.  (Dee didn't mention so explicitly,
but this seems to be a reference to the famous sign "Let none who have not
studied geometry enter here...")  Likewise, writes Dee (in his prolix style)
, this preface will warn the reader about the contents of the "Elements."
He then proceeds to a philosophical discussion of Mathematicks, whose
two primary topics are Number and Magnitude, Number being made of Units,
simple and indivisible, while Magnitude is arbitarily divisible, and has
length, width and breadth, ... etc.  The key reason for studying Mathematicks
is that numbers are beings of a third kind occupying a realm bridging the
Supernatural and the Natural, so that their study is uplifting.  To make this
claim more concrete and exciting he mentions one "Joachim," later referred
to as "Joannes Picus", who wrote a book of true prophesies, which according
to the "Duke of Mirandula" were obtained by the study of "Formal Mathematicks."
(I would be very interested to learn a bit more about this prophet!)  He
then discusses how, while the purest study is that of the integers, various
practical "reckonmasters" have developed fractions, radicalls and so on.
He gives one example that looks like the square root of pi, but the "pi" is
written in an odd manner - if that's what it is.  While not wishing to impute
any such base concerns to his readers, he promises no less than FIVE proofs
that the contents of the Elementa are quite practical.  Here we had to leave
off.

There were many odd things about this text which I will try to study more
carefully when Peter and I return to it.  Occaisional references are made
to the "Lion" or the "Lion's Claw", quite terse and unexplained, which I 
imagine have something to do with the picture of the lion at the beginning.
There is a reference to the "Allmighty Ternary," which is perhaps the Trinity,
but at the bottom of the same page -- curiously indicated by a small picture
of a hand pointing its finger! -- the Ternary is defined as the "Unity,
knot and Uniformity," an explanation unlike those I am familiar with.  
Written quite faintly in reddish ink, in a beautifully careful hand, in the
margin of one page, is the exhortation "Seek ye the kingdom of God and all
will be granted thereto".   All in all, quite unlike the geometry text I had
as a student.  :-)

I apologize for the fact that this has rather little to do with the Voynich,
save that it's an example of the mentality we may be dealing with.  From what
I read so far it seemed quite possible that the preface had an esoteric subtext
which only careful reading would disclose.  

-----
Peter Davidson raises the question dear to my heart, namely, whether
statistical analysis of the Voynich could ever determine it to be 
gibberish.  I feel that there are long-range (i.e., sentence- or
paragraph-length) patterns of word occurence in meaningful text that are
unlikely to appear in gibberish.  (Of course, if one finds such patterns
one can write a program to write gibberish having these pattern - but
it's unlikely that X (the Voynich author(s)) would have been so sophisticated.)
I would like to find out: "Given that a certain word appears at place
n in the Voynich, what is the probability that it appears again at place
n + m" -- and the same thing for other texts, such as the Book of Enoch,
medieval herbals, and so on.  Perhaps this or a more sophisticated statistic
can rule out some hypotheses concerning the Voynich.  Certainly the 
statistical difference between hands A and B and Jim Reeds' claimed oddities
about the first word in each line could, if studied carefully with statistics,
make it clear that this is not a word-for-word mapping of any "ordinary"
text.  

jb

From gauss!rand.org!jim%mycroft Tue Dec 31 11:18:12 PST 1991
Received: by gauss; Tue Dec 31 14:29:10 EST 1991
Received: by inet.att.com; Tue Dec 31 14:29 EST 1991
Received: from mycroft.rand.org by rand.org; Tue, 31 Dec 91 11:18:15 -0800
Received: from localhost by mycroft.rand.org; Tue, 31 Dec 91 11:18:13 PST
Message-Id: <9112311918.AA00297@mycroft.rand.org>
To: voynich@rand.org
Subject: Pronounceable Voynich text
In-Reply-To: Your message of Tue, 31 Dec 91 11:17:45 -0500.
             <9112311625.AA00207@rand.org> 
From: Jim Gillogly <jim@rand.org>
Reply-To: jim@rand.org
Date: Tue, 31 Dec 91 11:18:12 PST
Sender: jim%mycroft@rand.org
Status: OR


> reeds@gauss.att.com writes:
> It would be interesting to find the most pronouncible substitution 
> from Currier to phonetic values, using, as Kornai suggests, information
> about stops, glides, etc. I don't think we should transcribe in such a
> ``phonetic alphabet'' but it might well be worth our while to learn to
> pronounce Currier letters phonetically.

On Currier's original transcription page there's a string of equivalents
running down the left.  I applied them to a piece of f1r and got the
following:

   00101A VAS92.9FAE.AR.APAM.ZOE.ZOR9.QOR92.9.FOR.ZOE89-
    tehweths.ethdehn.ehk.ehlehm.wahn.wahketh.Qahkeths.eth.dahk.wahnteth-

   00102A 2OR9.XAR.O.R.9.FAN.ZPAM.ZAR.AR*.QAR.QAR.8AD-
    sahketh.Xehk.ah.k.eth.dehn.wlehm.wehk.ehk*.Qehk.Qehk.tehD-

   00103A 29AU.ZCF9.OR.9FAM.ZO8.QOAR9.Q*R.8ARAM.29-
    sethehU.wideth.ahk.ethdehm.waht.Qahehketh.Q*k.tehkehm.seth-

   00104A $OM.OPCC9.OPCOR.2OEOP9.Q*AR.8AM.OFAM.OE.OFAD-
    $ahm.ahliieth.ahliahk.sahnahleth.Q*ehk.tehm.ahdehm.ahn.ahdehD-

   00105A 2AT.9.SCAR.QAM.WAR.YAM#
    sehT.eth.wiehk.Qehm.Wehk.Yehm#

   00106A 98ARAIZO#
    ethtehkehIwah#

Sort of pronounceable.

Jim Gillogly

From gauss!UWAVM.U.WASHINGTON.EDU!MAX.U.WASHINGTON.EDU!RJB Tue Dec 31 11:34 PST 1991
Received: by gauss; Tue Dec 31 15:12:55 EST 1991
Received: by inet.att.com; Tue Dec 31 15:12 EST 1991
Received: from uwavm.u.washington.edu by rand.org; Tue, 31 Dec 91 11:35:48 -0800
Received: from MAX.U.WASHINGTON.EDU by UWAVM.U.WASHINGTON.EDU (IBM VM SMTP V2R1)
   with TCP; Tue, 31 Dec 91 11:35:41 PST
Date: Tue, 31 Dec 91 11:34 PST
From: RJB@MAX.U.WASHINGTON.EDU
Subject: Dee's Preface to Euclid
To: voynich@rand.org
Message-Id: <C112A16C40FFA0372F@MAX.U.WASHINGTON.EDU>
X-Envelope-To: voynich@rand.org
X-Vms-To: IN%"voynich@rand.org"
Status: OR

Dee's Preface has a modern edition, and either D P Walker or Frances
Yates has written quite a bit about it.

"Joachim" is not "Joannes Picus," unless I'm badly off base.  "Joachim"
is Joachim of Flora, who was a prophetic Franciscan abbott who
divided history (and religious dispensations) into the era of the
Father (Jewish), the Son (Christian), and the Holy Spirit.  Oddly
enough Joachim didn't get into trouble, but he became a major figure
in chiliastic and millenarian speculations of all sorts -- or rather
his tripartite historical scheme did.  Dee's own prophetic mission
with its "leading, had it succeeded, to a total reformation of all
kingdoms etc" or whatever Casaubon says, is clearly indebted to J of
F.

"Joannes Picus" is  J Pico della Mirandola, for whom again see DP
Walker  (_Spiritual and Demonic Magic from Ficino to Pico della
Mirandola_) and Frances Yates.

The sign that's "almost the sign of alchemical mercury" is probably
Dee's "Monad."  The canonical text is Dee's "Hieroglyphic Monad,"
of which two translations are available -- I think the academic one
was done by C H Justen, and published originally in _Ambix_.

That sign also shows up in the Chemical Wedding of Christian 
Rosycross, leading to a burst of Twilight Zone music and many
speculative volumes about who did what with whom, when.

(Typo alert:  that's "C H Josten.")

At any rate, Dee's interests were not unique, and he's certainly not the
only candidate for the creator of the VMs -- not to be confused with
VMS, which is in its own way quite hermetic...
--rjb

From gauss!math.mit.edu!jbaez Tue Dec 31 15:24:25 EST 1991
Received: by gauss; Tue Dec 31 16:01:42 EST 1991
Received: by inet.att.com; Tue Dec 31 16:01 EST 1991
Received: from WEYL.MIT.EDU by rand.org; Tue, 31 Dec 91 12:26:15 -0800
Received: from nevanlinna (NEVANLINNA.MIT.EDU) by math.mit.edu (4.1/Math-2.0) 
	id AA04596; Tue, 31 Dec 91 15:20:33 EST
Received: by nevanlinna; Tue, 31 Dec 91 15:24:25 EST
Date: Tue, 31 Dec 91 15:24:25 EST
From: jbaez@math.mit.edu
Message-Id: <9112312024.AA10828@nevanlinna>
To: voynich@rand.org
Subject: Dee's Preface to Euclid
Status: OR

Thanks for the info, Jim Reeds and RJB!  Identifying "Joachim" with Joachim of Flora and
"the Duke of Mirandula" and "Joannes Picus" with Giovanni Pico de Mirandola
seem sensible to me... I vaguely know these characters but didn't quite recognize
them in these disguises.  Dee's style is so florid and terse (if one can picture
this) that it's hard to tell whether he's identifying these fellows or not.

I will track down the references.  While Dee isn't the only candidate for X, he's
at least an interesting fellow.

jb

