We have two different semiotic systems.
The words say:
Tom went into the room.
The picture shows Tom entering a room. It also shows something the words have not mentioned: a dark shape beneath the bed.
The two systems have contributed differently.
But we do not leave their contributions separate.
When we read the words, we take the Tom in them to be the person shown in the picture. We take the room to be the room represented there. The picture, in turn, becomes relevant to how we understand what the words are saying.
One system has, in a sense, taken up the contribution of the other.
But this does not mean that the picture has interpreted the words, or that the words have interpreted the picture.
The picture remains a picture.
The words remain words.
What changes is the relation we construe between them.
The meaning-maker can relate the contributions of the two systems because they have been presented as belonging together.
This is why multimodal meaning is not simply a matter of adding one meaning to another.
Suppose the picture shows Tom entering the room and the words say:
Tom had never been here before.
The picture cannot, by itself, show us that this is the first time Tom has ever been there.
The words cannot, by themselves, show us exactly what the room looks like.
But when we take the two together, each changes what the other makes relevant.
The room we see becomes the room Tom has never visited before.
Tom's first visit becomes the visit we can see.
The contribution of one system becomes part of the meaning of the other.
This does not require the two systems to share a grammar.
It does not require one system to be translated into the other.
And it does not require some third semiotic system hidden between them.
What it requires is that their contributions can be related.
The reader does not first construct two completely separate meanings and then add them together.
The meanings are construed in relation.
That gives us a more precise way of thinking about what it means for different semiotic systems to work together.
They do not have to become the same.
They have to become relevant to one another.
But that raises a further question.
What makes it possible for the reader to treat these different contributions as belonging together in the first place?
The answer cannot lie entirely inside either semiotic system.
Perhaps we need to look beyond the systems themselves—to the situation in which they are functioning.
And that is where context enters the picture.