Basically, there’s a difference between what a language is capable (or not) of expressing (expressibility), and how easy it is to concisely express (certain) things in a language.
Or, as described by the user Stephen C on Stack Overflow:
expressiveness and expressibility are different things. For instance any algorithm you can write in Java is expressible as a Turing machine, but the language of Turing machines cannot be described as expressive, compared to Java. – Stephen C Feb 28, 2013 at 0:02
And, Wikipedia:
The term expressive power may be used with a range of meaning. It may mean a measure of the ideas expressible in that language:
- regardless of ease (theoretical expressivity)
- concisely and readily (practical expressivity)
Examples
RDF
The semantics of RDF are very limited, so it has limited expressibility: there’s a lot one cannot express in it. Note that a lot can be represented in it, which is precisely what it’s for. But without semantics superimposed on it by languages like RDFS or OWL, it’s very limited in that regard.
RDFS vs OWL
RDFS provides semantics which allow you to express yourself. It allows you to say high-level, conceptual things like “a is of type A”, and “b is a subproperty of c”.
OWL provides a lot more semantics, so you can express more in it, and hence it has greater expressibility.
Clojure vs C
Higher-order functions such as zip, map and reduce allow for greater expressivity, not expressibiity. Anything you can express in Clojure, can be expressed in C as well, but given its narrower semantics/vocabulary, it takes more effort to express it.