Before we can begin to assign individual sounds to individual characters we must first ask what kinds of sounds the language of the Voynich manuscript might have had. We cannot say exactly what sounds without knowing the language, but we can make some first steps based on language universals. There are rules for speech sounds which hold true across all or almost all languages that we can reasonably assume hold true for the Voynich language as well.
All languages have sounds (except sign languages, but they were neither well developed nor acknowledged in the 1400s) and the sound inventory of each is a subset of all the sounds humans can make. So, to begin with, we should not expect to find any sounds not found in other languages. Languages also typically abide by certain principles in the selection of sounds, with certain groups or kinds of sounds almost always present. We should, on the balance of probability, be able to assume that these groups will be represented, and can also speculate on the specific sounds they might contain.
Some of the following information has been gleaned from World Atlas of Language Structure (WALS), which provides a nice overview of linguistic features throughout the world’s languages. The data is from modern languages, not those of the 1400s, but the key points remain the same. We are interested in universals and tendencies which do not change.
Vowels: all languages have vowels, so we can be assured that the Voynich language definitely does. The minimum number of vowels is a bit hazy, but even as few as three is a common configuration, such as seen in Classical Arabic. Five is more usual, and WALS considers five or six to be average, so we should not expect a great deal of vowels overall. The main question about vowels is whether they are fully represented, as the influence of Semitic scripts could mean this is not so.
Consonants: although more will be said below about individual kinds of consonant, I just want to say a word here about how many there might be. The Voynich script is well known to be rather small, yet WALS proposes an average of 22 consonant sounds in the languages it lists. Many languages in the area of our greatest interest—Europe, North Africa, and the Middle East—have consonant inventories of average bigness or greater. Those languages with the smallest inventories, of fewer than fifteen sounds, tend to cluster in Australasia and America, which are beyond where the manuscript is likely to have come from.
But even if the number of consonants in the Voynich language is rather small we should still expect to see a spread of sounds. Languages with few sounds tend to maximize contrast so that different sounds sound different. There should be a few sounds each from many groups rather than many sounds each from few groups.
Stops: these are sounds such as /p, t, k/ and their voiced counterparts /b, d, g/, among other variations. All languages have at least a handful of stop sounds, and most in our target area have /p, t, k, b, d, g/. Some languages, such as Arabic and those influenced by Arabic, are missing /p/. But others have more stop sounds than those six.
Fricatives: the only languages without fricatives are far away from our area of interest, so we should expect the presence of fricatives. However, the number of fricatives in a language can vary widely. A minimum would be one or two sibilants, such as /s, z/, but ten or more is not impossible in our area. Included in this is the possibility of /h/, which is fairly common, though not always seen as a fricative.
The presence of affricates is also something of a wildcard: many language have none while others have half a dozen. It is impossible to guess whether an unknown language will have them.
Nasal: all the languages in our area have nasal sounds, as do most of the world’s languages. Typically /m, n/ are present, or some variations thereof. We should not expect a vast array of nasal sounds beyond this, as even /ng/ is not terrifically common.
Liquids: these are sounds such as /l, r/, along with a number of similar sounds. The lateral approximant /l/ is present in practically all languages in our area, with some also having lateral fricatives. Rhotic sounds like /r/ are also typically present, but there is much greater diversity and sometimes more than one such sound in a single language.
Semivowels: are sounds such as /y, w/, which we might expect one or more of to be present in the language. However, their obvious similarity to vowel sounds may make representation in writing unclear or indistinct.