While speaking on low level word structure I remarked that <l> appears in a particular way before <k> which made it an exception to the general model I was building. I proposed that it could be a digraph: a string of more than one character which is read to represent a single sound. For example, in English the ‘sh’ sound is represented by two letters. I would like in this article to expand on my thoughts about <lk> and more generally the idea that <l> makes digraphs with other characters.
The case for <lk> being a digraph is stronger than any other combination with <l>. The string appears 1,079 times in the text (though transcriptions may vary in the number), but it should not be seen simply as an expression of how the character normally acts before <k>. The character <l> is most similar in use to <r>, and next most to <d, s>, yet those combinations do not occur in any great numbers before <k>: <rk> has 19 tokens, <dk> has 33 tokens, and <sk> has 22 tokens. Something is unusual about <lk>.
Nor should the <l> of <lk> be seen as part of a prefix. Some researchers have noted words like <olky> and proposed that because over a thousand words begin with <ol–> it is a meaningful prefix. Yet hundreds of words also begin <or–, od–, os–>, and words beginning <lk–> alone occur 443 times. The string <lk> does not need to have <o> to its left, nor do other characters following <o> act in the same way. The presence of <o> before it is due to the higher level structure of a word.
It should also be noted that <lk> is mostly a feature of the Currier B language. It is roughly twenty times less common on A pages than B pages.
The presence of digraphs composed of <l> and other gallows characters is less secure. The string <lt> occurs 107 times, <lp> occurs 40 times, and <lf> occurs 39 times. Although <lf> is the least of the three its rate is actually rather great, being nearly 8% of all <f> occurrences, approaching the 10% for <lk> of all <k>. Even so, these number are still small and could easily be overlooked if not for <lk>.
Like <lk>, <lt, lp, lf> all appear at the beginning of words, and mostly occur in Currier B. They seem to work in the same way, even if less common.
Other digraphs, particularly <ld> (452 tokens) and <ls> (162 tokens), are possible. However, both are less obviously linked to Currier B language, and their distributions are unalike. The string <ls> seldom appears at the beginning of words, but two thirds of the time at the end of a word. The string <ld> occurs about 10% of the time at the beginning of words, but 55% in the word ending <–ldy>. It seems as though they have a different origin, and may simply be a valid string of characters with separate values, rather than a digraph with a single value.
The possibility of a digraph in the Voynich script is intriguing for three reasons. The first is simply that there may be more ‘characters’ than we thought, and that we should keep our eyes open to other digraphs. We have already seen this possibility with <e> and <i> sequences, but here two different letters have combined.
The next is that there may be more sounds than we thought, and the script’s relatively small character set may not be fully representative of the sound system of the Voynich language.
The third reason is that the digraphs only appear regularly on Currier B pages; but what can we make of this? Whatever sounds the gallows characters stand for it could be that they have two possible realizations which are not well distinguished in Currier A but are in Currier B. The two realizations should, naturally, be phonologically similar enough for some language to treat them as one, but also different enough for a language to also treat them differently. Something happens between A and B to make that distinction warranted, which is then made by adding an <l> before them.
It may be that sounds appear in Currier B which do not in Currier A, and it thus represents a different dialect or pronunciation. But it could also be that the language is the same but the writer wishes to make a new distinction between existing sounds. It could be that the Voynich language does not natively distinguish between certain sounds, but the amount of borrowings—presumably high in herbals—has forced the writer to change the way the script works. A similar thing happened in English where /v/ did not natively occur at the beginning of words and was not distinguished in writing from /f/, but the inbringing of so much French eventually made it necessary. This is a neat suggestion, but only that.
Although it is too early to speculate on what sounds <l> or the gallows characters represent, the presence of digraphs provides a further limitation on their possible identity. The ambiguous nature of the distinction they make also gives us a little insight into the sounds of the language.
I cannot at the moment think of a test by which we can prove <lk> and others are digraphs, but the possibility is intriguing and I believe it should be kept in mind. It certainly makes my model for low level word structure simpler, which I regard as helpful and positive.