Here are two definitions of comprehension from teacher-researchers:

Comprehension is the process of relating language to concepts in one's memory and to references in the real world. Comprehension is the sense of understanding what the language used refers to in one's own experience or in the outside world. Complete comprehension then refers to the listener having a clear concept in memory for every referent used by the speaker.
(Rost, 2002: 59)

Comprehension is not understanding what words mean, but is understanding what speakers mean. Even after getting the meaning of the words, the listener must still try to understand what speakers mean.
(Buck, 1999)

Both of these definitions imply that comprehension takes place when language links to our knowledge of the world. The idea is that through language, the listener is able to access the concepts and references which are in the speaker's mind.

However, that correspondence is never total. Words are not used to communicate meaning on their own, but in context with other words, which influence their meaning. Moreoever, speakers often do not explicitly communicate information which they assume is shared with the listener.

Much comprehension, therefore, is based on inference, with the listener filling in, interpreting what the speaker must mean. It is in this sense that the second definition stresses that comprehension is 'understanding what speakers mean'.

Usually before listeners can make inferences about the speaker's meaning, they have to make some sense of the raw material, ie the stream of sound. But even here, we use our background knowledge of the language system and of the world in order to 'fill in' partial information.

Comprehension then requires the integration of different levels of skill and different kinds of knowledge.