They could also be phonetic digraphs - double letters which are considered to be a single letter in the orthography.
Examples from modern languages would be the (old) double ll or the ch from Spanish, both of which were considered to be single letters in the alphabet; or ij in Dutch.
If we take k and ch to be distinct phonetic letters, then there could be a mixture of phonetics to allow the glyphs to be combined in a certain fashion. That also accounts for their regular placing if we assume Voynichese to be a transliteration of a language - or the phonetics being somehow beyond the encryption process. But given that all glyphs in Voynichese appear in regular patterns, we're only considering a subset of the language when looking at these bench glyphs.
Examples from modern languages would be the (old) double ll or the ch from Spanish, both of which were considered to be single letters in the alphabet; or ij in Dutch.
If we take k and ch to be distinct phonetic letters, then there could be a mixture of phonetics to allow the glyphs to be combined in a certain fashion. That also accounts for their regular placing if we assume Voynichese to be a transliteration of a language - or the phonetics being somehow beyond the encryption process. But given that all glyphs in Voynichese appear in regular patterns, we're only considering a subset of the language when looking at these bench glyphs.