Testing my own blog site (blog page) I noticed an assumption about name formatting that isn't universally correct

[{
  "key":"SZGR9VUU",
  "version":0,
  "itemType":"webpage",
  "tags":[],
  "title":"Rate Encoded Location",
  "websiteTitle":"AI stuff",
  "url":"https://jeblad.github.io/neural%20nets/2018/03/10/rate-encoding",
  "abstractNote":"Outline of rate encoded locations, how the brain might do it, and a possible approximation in an artificial neural net.",
  "language":"en",
  "accessDate":"2019-03-12",
  "author":[["John Erling","Blad"]],
  "source":["Zotero"]
}]

Notice the author field. At the site this is encoded as "John Erling Blad", but in the JSON structure "Blad" is split into a separate field. That is "Blad" is family name. In other languages this does not hold in general. For example, for w: Eggert Ólafsson the part "Ólafsson" is not a family name but a w:patronym. In such cases the name should probably not be split, that is it is not a simple (first name, family name) pair.

I have no simple quick-fix for this, except if there are no commas in the extracted names, that is they are not inverted, then they should probably not be split into (first name, family name) pairs.