Minims (2) and their importance in the Voynich script

Following my earlier post about minims – which you should read before reading this one – I now want to extend the debate to suggest that this area could have a major impact on our understanding of the Voynich script and signs.

Let’s start with a quiz: Look at these five signs and groups and try to say which Latin sound or letter they stand for. They are taken from just two words in a 15th century manuscript in the Wellcome library – not the Voynich manuscript, though curiously it was also once owned by Voynich.

minim table 1Minim quiz

How did you do 윤선생 앱 다운로드? Are they numerals? Or if they are letters, which ones? These are all made up of ‘minims’, and they illustrate one of the problems of reading Latin script from that period. The tricky part is that minims in different combinations can be used to make a variety of different letters, and we need two things to interpret them correctly:

a)      Context
b)      Knowledge of the word

Now, if I give you the context and meaning, the quiz becomes much easier – the original words are ‘Centaura major’ meaning ‘Great Centaury’, the plant analysed in my Feb 2014 paper, and the signs can be interpreted as follows:

minim table 3

centauraOriginal word

I think you’ll agree that unless you have the context and the word knowledge – which we don’t have with the Voynich script – it is a nightmare to tell if a letter is  ‘n’, ‘u’, ‘i i’, or in the case of three minims,  ‘i u’, u i’ , ‘m’ or even ‘w’ or ‘i v’. To some extent, in Latin and other orthographies of the 15th century, there were other clues to help, and some scribes added elements which helped to distinguish them, but the problem still remained windows 10 update.

Minims in the VM

Why is this important for the Voynich manuscript? Well, although some analyses give  38 signs or clusters in the Voynich script, with  22 common signs, some claim that this is very few for a ‘real language’  and that therefore the VM cannot encode a real language at all, at least in a sign-by-sign way.

Well, if we consider the above point about Latin carefully, and consider that the Voynich scribes might have adopted similar practices, and even borrowed the practice from Latin, the  result for our analysis would be to increase the number of possible significations for the V signs considerably 즐톡톡 다운로드.

This is because two very frequent clusters in the VM look like  ii and iii.  If the scribes were following contemporary Latin practice with minims, as seems possible, then these signs could in fact be intended to signify a wide range of different sounds/letters in different parts of the text. Two minims in one word might mean ‘ii’, or it might mean ‘u’ or it might mean ‘n’, and so on nisekoi 2nd term.

The example of the Wellcome Italian manuscript, roughly contemporary with the VM,  shows that this flexibility of signification, using the same minims, was very common. The reader needed to use content and word-knowledge to decide which one was correct.

What this flexibility means for the Voynich script, then, is that we might then have to rethink completely the ‘number of letters’ in the Voynich inventory. Instead of saying that it has just two clusters ii and iii, we would have to accept the possibility that they might, in different words, indicate a far wider range of sounds/letters, even as many as 14, not including ‘v’, as follows:

minim tablePossible range of signification for Voynich minims

Although to the modern eye this seems peculiar, it is clear from actual deployment of minims in mediaeval  scribal practice that this kind of flexibility did exist Chrome uninstalled download. If this analysis for the VM is possible, the implication is important – as I said, it significantly increases the range of possible readings of Voynich signs.

This could possibly add both consonant and vowel signs/sounds to the mix, though of course I am NOT saying that we will necessarily find Latin n, m u and i letters matching n, m, u and i sounds in the text.

Incidentally, this problem of how to deal with minims when transcribing the VM was recognised by Stolfi, obliquely,  in a posting in 2001 when he was talking about transcription meditation music mp3. He concluded with the question:

“What do you do with that? Nasty.”

However, apparently the ‘minim problem’, as we can call it, was so ‘nasty’ (as I agree it is) that it wasn’t dealt with in Stolfi’s transcription system or, it seems,  in later transcriptions such as EVA. That seems to me to be potentially a big problem, for reasons we can now consider.

Examples from the VM

So far this is just speculative, so let’s now look at some actual Voynich examples. Below are three words taken from the same Voynich page (2v, line 5, 6 and 1 respectively). The first has been transcribed (in the EVA system) as daiiin, the second as daiin, the third as  kooiin more help. Same two and three sign clusters, no?

However, if you look at them closely, there is a significant different between the ways in which the minims are joined. In Example One the last three, or maybe four downstrokes, including the last character transcribed as ‘n’, are clearly joined together, looking in fact like three or perhaps four connected minims and an upward flourish.

 daiin
Example One

In Example Two, by contrast, the minims are clearly separated, and also separated from the last character transcribed as ‘n’:

dain

Example Two

In Example Three, below, the two minims are clearly joined, whereas the last character (transcribed as ‘n’) is clearly separate Nahimic.

kooin

Example Three

It is easy to dismiss this as accident or chance, but that would be poor analysis. It is essential methodology to assume that what the scribes did had a purpose. The most likely explanation in this case is that they were conceptualising different sound sequences as they wrote, and therefore linking some minims and separating others to indicate different sounds, which by analogy with Latin could be i, u, n, m and so on.

This analysis raises an important problem for Voynich studies, because so far we have treated all ii and iii clusters as only two ‘signs’, in our counting, in our statistics and in our thinking 전설의 12장 다운로드. Our transcriptions – and therefore all of our computer analyses, entropy, Zipf and all the rest – have been based on that assumption. If it is wrong, we need a major rethink, and a major recount. Nasty indeed.

I would argue, then, that we need to take account of the possibility that these two minim clusters could hide a far wider range of signs, following standard mediaeval scribal practice with minims in Latin and in other languages. This then means that we need to re-analyse our Voynich transcriptions significantly, and maybe then our statistical analyses, in order to take account of exactly when the minims are joined and when they are separated 짱구는 못말려 게임 다운로드.

If we do not, then we run the risk of drastically misinterpreting the number of signs in the script as a whole, firstly. Secondly, this could potentially undermine every statistical test we run on the script as a whole, if we continue with transcriptions which potentially understate the possible number of signs in the script. Unless we act on this, we could continue bumbling around in the Voynich fog for a while to come.

References:

Good summary of the information on the V script: http://www.voynich.nu/writing.html

Herbal manuscript cited in the text:  Wellcome Italian herbal (Date: c. 1475 Origin: Italy, Language: Latin) (Once owned by Voynich) Wellcome Library London, MS.334

15 Comments

  1. I fully agree with the analysis in this post. Just commenting to add a funny example of this in medieval Dutch manuscripts: iiiiiiiie is how they wrote minne, the common word for love. Makes you wonder why they started dotting their i’s 🙂

  2. Derek Vogt

    Here is a boiled-down presentation of the thing about /a/, /al/, and /ar/, together with the thing about comparing only root words and leaving suffixes out of it. A lot of this isn’t directly related to EVA-i or what I said above about EVA-N, but I need to put it here to clarify that the /r/ I’m adding in these cases is not apparently a feature of “vowels” in general but just EVA-O, and doesn’t work for EVA-A or EVA-i as vowels, so the only way that it’s relevant to EVA-i is that it means EVA-N is not needed for /r/ and is thus free to be associated with EVA-i in the way I said it is.

    The representations for EVA-i and EVA-N in my otherwise phonetic transliterations here (a short stroke technically a dotless [i] or iota, for EVA-i, and a longer one, a backslash, for EVA-N) are just placeholders because I can not associate any sound with either of them.

    The format for each item in this list is…
    1. folio & EVA
    2. transliteration, with & without /a/ substitution & suffix separation
    3. sample cognate(s)

    66v okeodof
    akoatap; alkoat-ap
    alʔloat (Arabic)

    21r ofychey
    apnhon; arpnhon
    perpehen (Persian)

    21r pchofychy
    bhapnhn; bharpnhn
    perpehen (Persian)

    28r opchol
    abhaš; albhaš
    alabasa (Spanish)

    24v tchodar
    ghatwr; ghar-twr
    gaharu, agar (Indonesian/Malay, Hindi/Urdu/Bengali)

    20r kdchody
    kthatn; kthar-tn (or kthart-n)*
    satureia, $atrah, za’atar (Latin, Hebrew, Arabic)

    55v kcheedchdy oedain
    khoothtn aotwι\; khoothtn alot-wι\ (or alo-twι\)
    kurrâth al dubb (Arabic)

    06v koar
    kawr; kar-wr
    kharwagh, kharvay (Arabic, Armenian)

    39r tedochshd
    gotahxt; gotalhx-t**
    colchicum, qolkhyqum (Latin/English, Hebrew)

    *There is also another word in which it appears that “th” could be a digraph for some kind of sibilant, but I don’t yet know exactly how to depict that sound, so I leave [th] unchanged. Consonant order reversal “tr-rt” coud also be involved in this case, using the “t” which might or might not be part of the suffix.
    **This is a fun one because the vowel /a/ is not needed for this word, but it might have been the only way they could write the /L/ in. Also, there is an unrelated phenomenon that seems to happen in other Voynichese words with “oa”, which the “t” might have been inserted just to break up, which hints that it could be silent in this case.

    The above examples are cases in which the plants’ identifications and names were found while reading EVA-O as nothing but plain /a/. The tendency for /L/ and /r/ to seem to be missing right after it is a pattern that became apparent after enough examples like these had been observed. Once the pattern was known, it also helped make a couple of other possible connections that had not been apparent without it.

    38v okchop
    akhab; alkhar-b*
    al-kharshuf (Arabic)

    95v1 oekshy
    aokxn; alokxn**
    aldxan (Arabic)

    *apparent alternate name in the VM where another word (oteol:agoaš) had already been linked to that plant’s Greek name (agries)
    **not known to be a plant name, but is Arabic for a description which is built into the plant’s names in other languages

    Here is one example which seemingly could go either way; the connection to its numerous foreign cognates is apparent whether we treat this as /a/ or /al/, although I think the latter is a bit closer:

    41v keerodal
    kooratwš; kooralt-wš
    coriander, cilantro (English)

    Finally, the part where this becomes relevant to EVA-i and EVA-N: this reading of EVA-O is applicable to a few of the names in the original Voynichese phonetic paper from early 2014.

    03v koaiin
    kawιι\; kar-wιι\
    khartu, kharbaq

    29v koaiin
    kawιι\; kar-wιι\
    carum, karawaya, karo

    And that illustrates my two points about EVA-i (and EVA-N):
    1. These root words (along with 06v /kãwr/) have a potential counterpart for the expected /r/ already and do not need EVA-N or EVA-i to provide it.
    2. Removing suffixes to find cognates for root words cuts them out of possible phonetic analyses so far but is better for finding the roots’ cognates.

    The original impression that EVA-N was /r/ came from only these two examples and one other: kentaury. This word and its connections to cognates can simply do without the /r/, using /knt/ as the root word. However, the following EVA-A, even though I tend to think of it as the beginning of a suffix, does seem to have a tendency to show up after roots that end with a phantom /r/, such as these and 41v “koorãtwš” (coriandER, cilantRo), so I am also currently pondering a possible argument that a final /r/ (even a lost one) in the rood word causes the choice of a suffix starting with /w/ or simply gets converted to /w/ when followed by certain suffixes…

    • Stephen Bax

      Intriguing… I think I’m convinced! Do you want to write a short summary of your thinking which I can put as a separate post, to make the whole scheme clear and in one place?

  3. Derek Vogt

    Whatever EVA-i is or the Voynichese minims are, I’ve come to the conclusion that EVA-N is the same thing or one of them, not a separate letter.

    Before I can get to the observations that led me to that conclusion, the first thing I have to get out of the way is why I think it is not /r/ as originally proposed. I will get back to the relationship between EVA-N and EVA-i after the next few paragraphs.

    * * *

    Part of the case against EVA-N as /r/ is related to something it has in common with EVA-i: its unusual distribution, only occurring at the end of certain groups of letters which occur only at the ends of words but do so in a very large number of words, making those groups of letters appear to be suffixes. Although appearing only in suffixes does not mean it can’t be /r/, it does mean we can’t assign it that sound value or any other based only on words that use those suffixes when their foreign cognates include /r/ in the root words, because the Voynichese root word, the part that needs to be related to foreign cognates in order to demonstrate letters’ sound values, would not include the EVA-N or other letters of the suffix.

    For example, the plant on folio 14v has been identified as a betony, Latin betonia or betonica, and the Voynichese word starts with /_tn/, where the first letter (EVA-P), based on over a half-dozen other examples, appears to be /b/. We could arguably connect the fourth Voynichese letter there (EVA-CH) with English [y] and Latin [i], but even that could be pushing too far, not only because of interference with other examples of that letter but also because [i] could be said to be the beginning of the Latin suffix, meaning the root word ended at /n/. But the Voynichese word still has more letters left even after that one, which clearly can not be part of this root word, before ending with EVA-N. Using the last few letters in the search for cognates would be like looking for foreign cognates of English verbs including counterparts for our [ing], [ed], or [s] in the foreign root words, instead of looking for counterparts with just the roots that we attach those suffixes to, and doing it in this case would lead us to think the last few letters equated to Latin [ia] or [ica].

    So, what about the examples of EVA-N in the original paper that started Voynichese phonetics? There were three, and the /r/ in the root word(s) for two of them (which were spelled identically in the Voynich Manuscript) can now be easily accounted for in another way. They were /ka/ followed by an apparent suffix ending with EVA-N, and the given cognates were variations on /kar/. In my continuation of the original procedure of looking for comparable names for identified plants, I have encountered up to a dozen other examples of the letter for /a/ (EVA-O) correlating with other languages’ /al/ and /ar/. (Similarly, I’ve encountered at least five and possibly up to seven examples of /r/ correlating with Voynichese “oo” and “oa”, further establishing that Voynichese, at least in my expansion on the original phonetics, generally had a complex relationship with other languages’ /r/ which tended to involve what would otherwise be read as vowels.) Whether that means the sounds /L/ and /r/ were actually implied and pronounced in those cases, or had been lost, or were reduced but retained as some kind of coloring effect on the vowel, I don’t know, but I have no doubt that a connection of some kind here is real, which allows us to use /al/ and /ar/ as predicted sounds in the search for cognates of root words with EVA-O. That allows our two Voynichese /ka/ words to already be connected with other languages’ /kar/ just by the root words (the first two letters) alone, leaving us with no need to use the suffix for sounds that should have been in the root, no need for a second /r/ at the end, and no need for an awkwardly long conjectural bridge spanning the other letters between them.

    If this aspect of EVA-O had been known earlier, then presumably nobody would have even thought of the idea of going all the way to the end of the suffix and adding an /r/ there. In fact, if the plants & names had been worked out in a different order, including betonica above, we could now think we have two examples of EVA-N as something like /k/ or /q/, given that the /kar/ in one of these plants’ names was followed in some languages by “beck” or “baq”.

    Our only other EVA-N so far is in the word starting with /kntw/, followed by EVA-iN and EVA-iNY in its two appearance on that page, for kentaury. Considering everything after the /w/ (EVA-A), possibly including that letter itself, to be suffixes, we are left with just /kntw/ or /knt/ for the root word here, with no /r/ in sight. But this is already enough letters to make the needed connections without it, especially for a language with Voynichese’s unusual relationship with /r/. It could have simply been dropped, or converted or absorbed into the /w/ that we can see, or implied in some other subtle way I don’t know of yet. But this single word, with another murky multi-letter gap to cross before getting to EVA-N at the end just to use a letter in a suffix for a sound that was expected in the root, does not give us a sound value for EVA-N on its own.

    * * *

    So, if EVA-N can’t be said to be /r/, what can be said about it?

    Not only does it show up only in sequences that appear to be suffixes, but it also shows up only at the end of those sequences. And it only shows up after EVA-i or in a place where EVA-i would show up, such as after EVA-A at the end of a word. And despite nearly always being associated with EVA-i, it never precedes it. And despite being exclusive to suffixes, EVA-i is practically never the final letter, but EVA-N often is. In other words, EVA-N’s distribution throughout the manuscript is exactly the distribution that would be expected of “the last EVA-i in a group with no other letter after it”.

    A search at Voynichese.com yields several sequences transcribed as EVA-EN, but a close look at each of those cases individually reveals that the character transcribed as EVA-N is not the same thing as the one transcribed as EVA-N after EVA-i. Instead of a short straight stroke at the base, it has a little tight inward curve. In other words, after a series of EVA-i, you can get what looks like an EVA-i with a tail, and after a series of EVA-E, you can get what looks like an EVA-E with a tail, but they never trade places.

    Adding a curved tail, or bending or lengthening a stroke that was already there, is an established way to create terminal forms of letters in some Aramaic-derived alphabets, and we already have another likely example in Voynichese with EVA-R-&-M. In the case of EVA-i-&-E with tails, it would seem that instead of “end of word” or “end of statement”, the tail would mean “end of repetitions of this letter”.

    Finally, it’s mathematically clear that EVA-i’s relationship with EVA-N has to be fundamentally different from its relationships with other letters. Several other letters are common at the ends of words, including after at least one EVA-i, but differences show up in how often they follow particular numbers of EVA-i. In each line below, the numbers will show how many words end with that letter alone, then how many end with that letter preceded by one, two, or three EVA-i:

    -y: 15258, 4, 1, 0
    -r: 5595, 591, 127, 0
    -L: 5829, 19, 7, 0
    -s: 1058, 25, 23, 0
    -m: 987, 45, 17, 1
    -n: 149, 1721, 4042, 147

    For the others, the highest number comes first and they decrease with each step after it. That’s normally to be expected for any letter; singles are more common than doubles, which are more common than triples. But EVA-N doesn’t follow that pattern. First the numbers increase and then they decrease. The decreasing part can be attributed to the same cause as with the other letters, but something else needs to be the cause of the increasing part. Part of the eplanation could be that EVA-N is just what EVA-i looks like when the suffix it’s in doesn’t have another letter coming after it.

    • Stephen Bax

      Thanks Derek, an interesting set of ideas. So can I recap and you correct me? 🙂

      -you feel that the /r/ sound is in places inherent in the vowel, so that where we have for example /kau/ we should read it with a final /r/ due to ‘colouring’?

      -you therefore feel that the final flourish which results in the EVA:N character merely signals the end of that group of vowels?

      • Derek Vogt

        The only vowel I’ve seen correlating with a vowel followed immediately by /L/ or /r/ in other languages is /a/, EVA-O. Two possible interpretations for that are that the /L/ or /r/ was pronounced but implied instead of being written with its own separate letter, and that the sound was entirely lost in these instances in Voynichese. The idea of the vowel being R-colored or L-colored is a third option, a possible middle ground between those extremes.

        I don’t know of a way to predict at this time which option actually describes Voynichese pronunciation. But, for identifying cognates in other languages, that isn’t necessary. All that’s necessary is to keep in mind that the cognates we’re looking for seem to sometimes have /al/ or /ar/ in EVA-O’s spot instead of /a/ alone. So, for the three words on folios 3v, 6v, and 29v that would seem to start with /kaw…/ or /kau…/ (EVA-KOA), the relevant secondary reading that would give us is /karw…/ or /karu…/, with the /r/’s place being before EVA-A (/w,u/).

        I hesitate to actually write that [r] in my transliterations because I’m not sure the /r/ was pronounced in Voynichese and I prefer to transliterate one character to one character as much as possible, but writing it does give the best idea of what the cognates involving this phenomenon should look like. In my own files at home, I mark these spots not with [al] or [ar] but with [ã], which can be thought of a bit like the French [^]: silent but indicating where a subsequent sound was before it got lost. So I have these particular words starting with [kãw…], where [kã] alone is the root word equating to other langauges’ “kar”, and [w] is just the first letter of the suffix.

        An effect of cutting the suffixes out of the phonetic interpretation is that EVA-i is not used in any of the sounds we can link to any cognates so far. It only shows up at the end, after the party’s over and the words have already been matched up based on the sounds of their earlier letters. And EVA-N is the same way. Not only does this give us no hint of the sound value for EVA-i or EVA-N, but other factors make me suspect it/they didn’t even have any. (I’m thinking of grammatical ideas that come in small groups of simple categories, especially those that can be thought of as sequential, like first person, second person, third person; or subject, direct object, indirect object; or singular, dual, triplural.) I know it’s weird to contemplate non-phonetic symbols with purely grammatical significance in an otherwise phonetic writing system, but what language has something in suffixes, with a huge tendency to repeat itself, that doesn’t appear in root words or prefixes and never occurs in two separate groups in the same word with another letter/sound between them? At least uniting EVA-N with EVA-i means we only have that problem with one letter instead of two.

        So, although I do believe EVA-N is the last EVA-i in a group (at least when the same suffix has no other letter following instead), I can’t call it the end of a group of “vowels” as your question put it, because I see no connection between EVA-N and anything else other than EVA-i, and because I can’t even call EVA-i a vowel. (And if it’s a minim, then groups of it could be a vowel sometimes and not at other times.)

  4. Stephen Bax

    Just to add to this – if you look at the ‘quire number’ at the bottom of f70v1 you will see the number 10 plus three minims, plus the Latin abbreviation for ‘us’.

    See here:

    http://www.jasondavies.com/voynich/#f70v2_f70v1/0.926/0.977/5.00

    This is clearly intended to mean ‘Decimus’, i.e. the end of the tenth quire. See here for explanation of the quires.

    The M here is written in three minims, in exactly the way I describe in this posting above. Although this is probably by a different scribe, it is still good evidence of what I am arguing here, that the minims in the manuscript could be read in a large variety of ways.

  5. apoplaws

    Your work is truly fascinating and interpretation of the characters and words is very convincing. Also the theory of small and later disappearing culture seems to be plausible.
    I believe, one thing related to this nasty minims should be done. If the evidences against meaningful content of the VM based on some probabilistic or information theoretic measures are to be refuted because of minims, one should calculate maximal possible difference from such a measures taking over each possible assignment of the actual characters to minims groups and check if newly obtained measures are within the acceptable margin. In fact if they are it may give some clue what the correct assignment of the characters to minims’ groups is most probable.
    This seems to be pretty mechanical thing if some of the transcription takes the fact of the existence of minims into account and marks characters or groups as built of minims.
    If not, this seems to be hard because requires creation of the new transcription. However it still may be not as hard at seems at the first sight. I think it does not require reviewing whole the manuscript again. The minims were interpreted as some special characters or groups of characters already , so looking into this suspicious groups in the existing transcriptions should point us to manuscript itself and allow to find real minims. Exactly as if we have transcription of manuscript in Latin alphabet and look for the place where in the transcription are such a letters or groups as “u”, “i”, “ui”, “n”, “m”, “ni” etc.
    For task of estimating new measures and identification of minims help from computing community would be beneficial.

  6. Carl

    By keeping your research in mind and the links you draw to both the greek, middle-eastern (perhaps Persia), and India and the lack of written history my first thought was of the Roma`s or commonly known as gypsies. I couldnt find the map with the linguistic mapping of their language but it shows a strong connection to your thesis. http://dnaconsultants.com/_blog/DNA_Consultants_Blog/post/Gypsy_Migrations/

  7. bdid1dr

    Prof. Bax, I am familiar with minim’s from my pre-med studies. However, I’d like to explain my translations of several combinative scripts used throughout Boenicke Ms 408. I’ve tried to demonstrate for persons who may not be familiar with pharmaceutical terminology, much less the Latin terminology which is written in the script of whichever nationality was producing manuscripts. My favorite B-408 alphabet letters are ‘m’ and ‘n’. I describe them as ‘looking like’ fish-hooks’ — one barb for ‘N’ and two barbs for ‘M’. So, imagine the fun I could have with the word ‘imagine’. Just hold your right hand up to your eye level as if you were trying to form a backward-facing letter ‘C’. Your hand will form the shape of the alphabet ‘N’. To form the alphabet letter ‘M’ , clap your hands with the left palm slightly lower than the right, so that the thumbs appear as two barbs.

    So, if one combines the numeral 8 figure with either of the barbed letters, one can read either ‘aes-an’ (one ‘barb’) or ‘aes-am’ (two barbs). Call the ‘barbs’ bumps, if you like, because one scribe seems to be writing at a greater speed than the other.
    I first began translating with a letter (written in Latin) which was a prescription for making colloidal silver. The letter is in Phillip Neal’s huge ‘Kircher’ correspondence files. Kircher apparently was horrified by the letter’s content; because he wrote in a very shaky hand, one alpha-character on each line of the note a-l-c-h-e-m-y.

    • Stephen Bax

      I’m not too clear what point you are trying to make here, exactly?

      • bdid1dr

        My point: The word ‘minim’ can be written in “Voynich” script with only 3 characters.

      • bdid1dr

        another good word: mo-mn- tm

  8. I think your analysis here is on the right track. From the transcriptions I’ve done of fifteenth century Middle English cookbooks, I know how hard it can be to distinguish between n, u, and m, and that’s when I have plenty of context.

    Another thing ocurred to me (and which you’re likely already aware of) that would throw off letter frequencies is the use of single letter prefixes. I don’t know if it’s applicable for V, but languages like Russian often mark noun case with a single letter prefix (e.g. V or F).

    Lastly, and this is pure speculation, I keep thinking that the original language may be related to Farsi. One issue with that is that the V Ms is obviously written left-to-right and Farsi is the opposite. Many of the letter forms though look similar to those in the Pahlavi script, only reversed. This makes me wonder if V was a 15th c. attempt to copy or rework a Pahlavi document to be more western.

    • Stephen Bax

      Thank you for your insightful and though-provoking suggestions. Let me take the points one by one:

      1. Minims: thanks for your support. That is certainly my experience of mediaeval manuscripts.
      2. Prefixes: no, I didn’t know that in fact. Very interesting. I will explore it further. Thanks for the insight.
      3 Persian: I must say I feel there may well be a Persian link, maybe partly in the script but maybe also in the underlying language. I feel this might be in terms of em>borrowing from Persian however. I don’t think the underlying language is Persian, but it could be!

      Thanks for your contributions. I’m grateful.

Leave a Reply to bdid1dr (Cancel Reply)

Your email will not be published or shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

*