I agree that Scheme is not âpurely functionalâ. However, given that the term was used here in reference to the use of Scheme, one could assume that the other common, perhaps less well defined, use of the term was implied, meaning a language whose syntax and functionality makes it convenient to use functions as arguments and return values from other functions (as e.g. is done in Schemeâs Wikipedia article, this blog post, or in The Jolly Writer,Sec. 14.1.3). In this sense, it is not a matter of putting langues in âcasesâ and I understand that languages can support multiple paradigms (as many languages do these days). I would then also say that TeXmacs relies heavily on many paradigms: functional methods, contextual overloading, macroâs and perhaps more.
Using this definition, I donât see any evidence for statements that itâs the functional paradigm that would be turning people away from certain languages.
By the way, by âmultiple dispatchâ, I would understand this Julia-style behaviour in TeXmacs:
(tm-define (square x)
(:require (number? x))
(* x x))
(tm-define (square x)
(:require (string? x))
(string-append x x))
Scheme] (square "hello ")
"hello hello "
Scheme] (square 2)
4
while TeXmacsâ contextual overloading mechanism seems much more general than this.
Anyhow, thatâs just a matter of terminology, which is maybe not that interesting to discuss 
What is more interesting to me is what properties make Scheme so uniquely suited to defining a new programming sub-language. What functionality is needed for that? Wouldnât this be possible using Juliaâs meta-programming features as well?