Positional Rigidity in the VMS

[Emma May Smith]

So, the first thing we need to distinguish is between script and sounds. Most scripts map onto sound only imperfectly and we don't know how good or bad the mapping is for the Voynich script. We have to presume that Voynich characters (or what we think might be characters) represent a single sound, even though that might not be true. However, when we discuss other languages we need to be clear when we're referring to the script and when we're referring to sound. I'll write Voynich script in square brackets [] and sounds in slashes //.

Next, we need to distinguish between words and syllables. Many words are made up of more than one syllable, and the same is true of Voynich words (we can not only presume this, but propose this with proof). Sounds are structured within syllables in all languages, and we can broadly think of syllables having three parts: onset, nucleus, and coda. The nucleus is the vowel, the onset is all the consonants before the vowel, and the coda all the consonants after a vowel. The onset and the coda are often optional, but the nucleus is obligatory. Syllables can only have one vowel (at most a compound vowel).

Most languages have different rules for what consonants can appear in the onset and coda, and how they are structured. Notably, the rules for the onset and the coda are not the same in many (or most) languages. Typically, onsets are more complex than codas, but that's not always the case. There are some general linguistic rules about how consonants are ordered when more than one appears in either the onset or the coda of a syllable.

Also, we must be aware that whole-word rules may determine where sounds occur. Some sounds can occur at the start of a syllable, but only if they are at the start of a word.

Let's get on to the issues with Voynich words and how they are structured. Specifically, we want to know if sounds in natural language can do what Voynich characters do.

1. Constraining characters to the start, middle, or end or words.

The character [q] is the most obvious example of this. It almost always occurs at the beginning of a word. Is this natural language like? Yes. The sound /h/, for example, was almost always word initial in Ancient Greek. The sounds /ng/ can only be word initial in Maori. Many languages, English included (some dialects), only allow a glottal stop at the beginning of a word, though it's not really phonemic.

Some characters, such as [t, k], may appear only at the beginning or in the middle of words. The reason for this is that the end of a word is also either the nucleus or the coda of a syllable. If [t, k] aren't vowels, then they would have to appear in the coda. As many languages restrict certain consonants in the coda, this is no surprise. Something like one in ten languages don't allow any sounds in the syllable coda, and most have other restrictions.

2. Constraining character adjacency.

I'm not sure exactly which characters you are speaking of here, so I'll be broad. Obviously some characters written as two letters in our transcription are likely to be one character in the script, such as [ch, sh]. Most of the other rules for character adjacency depend on how syllables are structured.

When we have more than one consonant in the onset or coda of a syllable they are typically ordered according to their sonority. We won't worry about what sonority is, but it is enough to say that sonority should be highest toward the vowel and lowest further away. So in any language /kra/ is more likely than /rka/ and /ark/ is more likely than /akr/. There are always exceptions, but it is a general tendency.

Thus we might find [e] always before a vowel. I don't know what [e] stands for, but it may be that it has the highest sonority of any consonant and thus must be followed by a vowel. The character [ch] must have a lower sonority because [cheo] is an acceptable sequence but [echo] typically not.

We must also remember, as mentioned above about coda restrictions. Characters such as [l, r] might often be preceded immediately by [a, o] simply because 1) they're in the syllable coda, and 2) only a single consonant is allowed in the coda. It's not possible for them to be further away from [a, o] than this unless they're in the onset, in which case they would be preceded by a space because sequences with them in the second position in an onset don't seem to be possible.

3. Constraining character repetition.


If the sounds in a language don't repeat then there is no reason for the characters to do likewise. Most repeated characters in English are not representing repeated sounds. 'Glimmer' has one /m/, 'less' has one /s/, and 'muddle' has only one /d/. Some repeated vowels, such as 'oo' and 'ee' in England are representing long vowels, not repeated ones.