The "top-level" structure of JSON is usually an object, but it can be a list.
> JSON fundamentally IS made to represent tabular data
No, it's really not. It's made to represent objects consisting of a few primitive types and exactly two aggregate types: lists and objects. It's a textual representation of the JavaScript data model and even has "Object" in the name.
> Why make it able to represent tabular data if that's not an intended use?
It's mostly a question of specialization and ergonomics, which was my original point. You can represent tabular data using JSON (as you can in JavaScript), but it was not made for it. Anything that can represent """data""" and at least 2 nesting levels of arbitrary-length sequences can represent tabular data, which is basically every data format ever regardless of how awkward actually working with it may be.
In the same way, there are hierarchically structured datasets that can be represented by both json in hierarchical form and csv in tabular form by repeating certain variables, but if using csv would require repeating them too many times, it would be a bad idea to choose that instead of json. The fact that you can do sth does not always make it a good idea to do it. The question imo is about which way would be more natural, easy or efficient.
The reverse is true as well: being more specialized is a description of goals, not advantages.
The big advantage of JSON is that it's standardized and you can reuse the JSON infrastructure for more than just tabular data.
I'd definitely put that in my list of falsehoods programmers believe about CSV files.
> This header will contain names corresponding to the fields in the file and should contain the same number of fields as the records in the rest of the file
But of course, CSV is the wild west and there's no guarantee that any two encoders will do the same thing (sometimes, there's not even a guarantee that the same encoder will do the same thing with two different inputs).
Headers should have as many rows as possible that contain data items for their column and data items in a row should have a header for the respective columns, but real CSV files should be assumed to have incomplete or variable length lines.
JSON is lists of lists of any length and groups of key/value pairs (basically lisp S-expressions with lots of unnecessary syntax). This makes it a superset of CSV's capabilities.
JSON fundamentally IS made to represent tabular data, but it's made to represent key-value groups too.
Why make it able to represent tabular data if that's not an intended use?