Paragraph 3
It seems pretty clear to me that this paragraph begins with 'pour empecher une femme de...' (to stop/prevent a woman from...) so I tried to guess what a pirate might want to stop a woman from doing. My first thoughts were 'getting pregnant' or 'being unfaithful' but these didn't seem to fit the cryptogram too well. I say that despite the fact that many people see the sequence 'DOBAUCHE' and assume that this corresponds to débauche, débauché or débauchée. 'Débauche' is French for 'debauchery', meaning excessive indulgence in sex, alcohol, or drugs, just the kind of behavior you might expect of pirates and their women! But I reject that interpretation because I think my interpretation works better in terms of the symbols of the cryptogram.
A little further in the text is the sequence 'LOREIL' which many people assume is a mis-spelling of 'l'oreille' meaning 'the ear'. I reject that too, but propose a related word. I think the sequence 'LOREILJ' is a slightly mis-copied 'LOREILE', and that La Buse actually meant 'l'oreillé', or more correctly 'l'oreiller' which means pillow. Having made this guess I looked back and saw the sequence 'REN' and then it clicked: another thing that a pirate might want to stop a woman from doing is snoring! To snore in French is 'ronfler' but La Buse, being a poor speller, might well have written 'ronfler' as 'renfler' or 'renflé' since these spellings are phonetically equivalent to 'ronfler'.
I did some research and was pleased to read that one traditional method to stop people snoring is to sprinkle certain essences or oils onto the person's pillow. This site, for example, suggests putting a mixture of essential oils on the pillow, including orange oil. Some sites recommend olive oil as a useful aid for sleep, but usually as something to be swallowed rather than placed on the pillow. However, oily olives ('olive grasse', in French) can be used in soaps and therefore mixing them with orange essence might make a good sleep-inducing formula. It maybe also that the orange essence will dissolve in the olive oil. In this book (published in 1845) it says 'Huile d'olive grasse ou commune faite avec une espèce d'olive petite et maigre. L'huile extraite de cette espèce qui vient sur des arbres de haute futaie est utilisée pour la fabrication des savons et aussi pour l'ensemencement des laines et la préparation des draps' or 'Oily olive or common olive oil made with a small and lean olive. The oil extracted from this species which comes on high-growth trees is used for the manufacture of soaps and also for the seeding of wool and the preparation of sheets.' If olive oil can be used in soaps and to prepare sheets then putting some on a pillow can make sense. When I saw sequences of letters in the cryptogram that resembled 'oranges' and 'olives grasses' I felt even more certain that this is indeed the correct sense of this paragraph.
Ignoring as usual the characters that seem to have been included as clues for the next layer of the cryptogram and which can all be expressed as digits as above, we have:
Pour empêcher une femm de renflé, vous n'avé qu'à fouetter des oranges et d'olives graces por épandre que l'uile sur l'oreilé.
In modern French, and correcting errors:
Pour empêcher une femme de ronfler, vous n'avez qu'à fouetter des oranges et des olives grasses pour épandre que l'huile sur l'oreiller.
In English:
To stop a woman from snoring, you just have to whip together some oranges and oily olives to then spread just the oil on the pillow.
Notes:
The cryptogram text includes the sequence 'FEMME' but in paragraph 5 La Buse spells woman as 'femm'. I will assume that he is consistent in that incorrect spelling, which means that I have to interpret the 'e' after 'femm' as another of the added symbols that are clues for the next layer of the cryptogram. Furthermore, since it cannot be a coincidence that all these added symbols are of the ambiguous type that can be interpreted as either letters or numbers, I am inclined to assume that this symbol represents a '2' rather than an 'E'.
La Buse has once again made a spelling error by writing 'uile' instead of 'huile' which is phonetically the same.
In the interpretation above I suggest that La Buse wrote 'des oranges et d'olives graces' where we would expect 'des oranges et des olives grasses'. That includes the suggestion that La Buse once again gave a spelling that is wrong, though phonetically correct, writing 'grasses' as 'graces'. But it's also possible that he wrote 'grasses' which would mean the spelling was correct but also that the correction is less faithful to the original symbols. It would also mean, and this is the crucial point, that there would not be the 'embedded symbol' in this word that appears in the above interpretation as a red '3' (I put the embedded symbols in black if I'm confident that they are present as shown, and in red if I am less confident).
Looking at the colored chevrons that indicate how close my corrected characters are to the originals you will see quite a number of orange and red chevrons which indicate that the copier must have made large and perhaps rather implausible errors when copying the symbols. This does bother me, but at the same time we don't know in what condition was the original document. If the original document was very small, dirty, badly drawn or damaged then serious copying errors become plausible.
One word in particular stands out as unsatisfactory: RENGT becoming RENFLE. That's very unsatisfactory because I had to add an extra letter that has no corresponding symbol, something I did nowhere else in my entire interpretation of the cryptogram except for a character that I think was obliterated by the 'large dot' on line 15. However, I do not have a better solution for this word at this time. If I can find a better interpretation then I will of course update this site...
At the same time, if I worry because I have to add a character or two doesn't that indicate that my interpretation is much more faithful to the original text than all the other interpretations that are in circulation?
Another unsatisfactory element is the word 'POR' which I take to mean 'POUR'. But La Buse spelled this word correctly elsewhere in the cryptogram so why would he make this clumsy error here? Did the copier fail to copy one of the letters? With more than 500 symbols to copy, all meaningless to the copier, one would not be too surprised if the copier were to make errors like this.
The sequence 'LAD' becoming 'DES' also works very badly in terms of the symbols and could therefore be wrong...
In the first word of this paragraph I do not mark 'POVR' as an incorrect form of 'POUR' because in old French the letters 'U' and 'V' were more or less interchangeable.