Custom Data Source API: how does the Filter object get used recursively?

Looking at the current documentation for the Custom Data Source API in order to implement it in my project (https://www.flexmonster.com/api/filter-object-for-requests/). The Filter object is defined recursively:

[{
    "field": FieldObject,
    "include"[]: {
        "member": string | number,
        "filter": FilterObject
    },
    "exclude"[]: {
        "member": string | number,
        "filter": FilterObject
    }, 
    "query": {
        (condition): string | number | string[] | number[]
    },
    "value": {
        "field": FieldObject,
        "func": string
    }
}]

Filter and Filter.include[0].filter have the same schema/are the same type.

It isn't made clear anywhere how this is actually used, and the only explanation we have to go by is that this nested filter "specifies which filters should be applied to the member."

What is a use case for the nested filter property in either the include or the exclude properties of the filter?