Several notable cases in point come to mind: I mentioned the UK, where cognatic grandchildren of a monarch -- even if reigning -- are not normally styled with the princely title. As such, it required letters patent by King Edward VII in 1905 and King George VI in 1948, respectively, to grant/ensure this for the offspring of their eldest daughters. The Prince of Wales may have been born within the lifetime and reign of his maternal grandfather, and to a mother who was only heiress presumptive (not apparent) to the British throne. But it would have been too great an indignity for him to be styled merely by the courtesy title Earl of Merioneth.
In the Netherlands (as pointed out), Queen Juliana (while still enthroned) had to do this not only for the prospective children of Princess Margriet (who until 1967 was second in line to the Dutch throne) but also, those of the then-Crown Princess Beatrix. It was made clear that the children of Beatrix would be styled as Prince(ss) with the qualification of Royal Highness, while those of her sister would be Prince(ss) with the qualification of Highness.
If it came to that, I believe that there would have been a royal decree from Queen Wilhelmina, regarding the titles and styles of the then-Crown Princess Juliana, since (once again) cognatic grandchildren would be born to a reigning monarch.
By contrast, no decree was required for Juliana herself, since she was born to a queen who was already enthroned (as opposed to being merely heiress). As with agnatic grandchildren, nobody questions that the children of a sovereign parent are automatically styled as Prince(ss), with the qualification of Royal Highness (I'm talking about kingdoms, of course).
But in 2002, the opposite happened: Queen Beatrix issued a decree to DENY the princely title to the prospective first child born to Prince Constantjin. Otherwise, my understanding is that Countess Eloise would have automatically been a princess at birth, with the qualification of at least Highness. It has been virtually an ancient and universal tradition that agnatic grandchildren of a king or queen regnant are styled as princes and princesses -- as long as they are born to approved marriages.
Another good example of stripping the princely title, as a result of issuing Letters Patent, would be the well-known case in 1917 of King George V of Great Britain. Both his Coburg and Hanoverian cousins lost their titles of Prince(ss) of the United Kingdom as a result. For until that point, they had followed German tradition, according to which titles and styles are unlimited in dynastic male lines. As it was, the king not only stripped his German cousins of their British royal titles but also, restricted the said titles even within the British royal family. Accordingly, Alistair of Connaught was deprived of his princely title, since he was only a great-grandson (albeit in the male line, styled only as Highness) of a monarch.
So whether one is expanding or restricting the use of titles and styles, it seems that the purpose of letters patent (or decrees) is to go CONTRARY to law or tradition. Is this correct?
Now the Belgian case in 1991 had nothing to do with any royal decree: giving the titles Prince(ss) of Belgium to Astrid's children resulted from the constitutional amendment which changed the succession law. But the situation with their father in 1995 was different: Lorenz was granted the title Prince of Belgium by their grandfather, who in 1993 had succeeded to the throne as King Albert II. So his case did result from a decree.
As did the case of the present Queen Mathilde, who in 1999 married a crown prince who had not yet succeeded to the throne. If I understand correctly, Belgian law in 1991 stated that the wife of a prince was not automatically a princess. As such, she had to be created into one.
So while the law took care of Princess Astrid's children, her husband and sister-in-law had to be given Belgian royal titles through decrees.
But this fact only underscores my observation about letters patent and decrees. Is all this correct?