Paragone’s claim is that languages “program” what we think; like in computer applications, they’ll only behave in the way they were programmed, you won’t see e.g. an image editor being magically able to edit text files, because it wasn’t programmed for that. That’s already in strong Sapir-Whorf territory.

In the meantime, all your examples point to weak Sapir-Whorf: languages influencing what we think.

That’s why I compared the situation with walking vs. streets. Sure, you might go further if you walk through a street than if you do some weird parkour across a city, so having streets or not does have some influence how you walk. But you don’t need streets to do something as basic as walking, you know?

So, like. What you are saying is perfectly reasonable, but it’s different from what Paragone is saying, you know?

---

And, when it comes to social vs. grammatical genders, languages with the later always make such a mess of it that it’s hard to conclude anything about its impact over the former. And that’s by necessity; noun classes (grammatical gender is a type of noun class) are only useful if all nouns are dumped in some category, even if that wouldn’t make semantic sense. Like “clock” in the masculine but “table” in the feminine, or words for human groups of mixed social gender using the grammatical masculine, or the word for “person” being using the grammatical feminine.

If anything, I think it’s more useful to look for the opposite: how society shapes language, not how language shapes society. Because it’s well-known languages are mutable, and social pressures can lead them one or another way. And you do see a lot of pressure in multiple languages (some with grammatical gender) to use more words and expressions that leave social gender ambiguous.

For example, in Portuguese I’m seeing more and more tests recruiting actors being called “teste de elenco” (cast test) instead of “teste para atores” (lit. test for actors). “Elenco” (cast) is as grammatically masculine as “ator” (actor), but unlike the later it doesn’t imply anything about the social gender of the people being potentially hired.

Another example (still from PT for my own convenience) is that I’m seeing a lot of strategies popping up to counter the “default masculine for mixed groups”; like this.

• [More traditional] Alguns acharam o vídeo chato.
• [Word-by-word translation] Some (M) found the (M) video (M) boring (M).

“O” (the) and “chato” (boring) are masculine there because “vídeo” (video) is masculine, so it triggers agreement, but that is not a big deal. Focus on the “alguns” (some) instead; it’s a masculine word being used to refer to a group of indeterminate or mixed social gender, so some people

So a sentence like this nowadays often has that “alguns” replaced. Like this:

1. Algumas pessoas acharam o vídeo chato. [Some people (F) found the video boring.]
2. Muita gente achou o vídeo chato. [A lot of people (F) found the video boring.]
3. O povo achou o vídeo chato. [The folks (M) found the video boring.]
4. Acharam o vídeo chato. [Ø found the video boring.]

#1~#3 work like that “elenco” I mentioned above; you’re sticking to the grammatical gender, but nothing assumes the social gender. #4 uses an indeterminate subject, so you don’t need any word there.

There’s also some more iconic strategies, such as saying “algumes”, “algum@s” or “algumxs” instead of “alguns”, as if Portuguese had a third grammatical gender associated with some human beings. But if I had to take a guess they aren’t as popular as the above, and a lot of them only work for writing, not for speaking.