In Critical Discourse Analysis, you explored how language choices build socio-cognitive models that shape our understanding of the world. Visuals work in a similar way. They don't just show reality; they interpret and represent it. This is the core of the ideational metafunction in social semiotics: how images represent objects, people, and their relationships.

Every visual text presents a version of reality. To understand how, we can sort these representations into two broad categories: narrative and conceptual. Narrative visuals tell a story, depicting actions and events. Conceptual visuals explain or classify, showing us what something is or what it's made of.

Narrative patterns in images show us a world in motion. They represent actions, events, and processes as they unfold. The key element to spot in a narrative visual is the vector—a line, either visible or implied, that connects participants and signals an action.

A vector can be formed by an outstretched arm, a pointed finger, the direction of a gaze, or even a tool being used. It creates a dynamic, a sense of something happening. These narrative processes break down into two main types: actional and reactional.

Reactional processes are about seeing. The vector here is the eyeline or gaze connecting a participant to what they are looking at. The one looking is the Reactor, and the person or thing being looked at is the Phenomenon.

Like actional processes, these can be transactional (a character looking directly at another character) or non-transactional (a character gazing off-screen, lost in thought). In both cases, the gaze directs our attention and tells us who or what is significant in the narrative.

When an image isn't telling a story, it's usually explaining a concept. These are conceptual structures. They lack vectors and instead present participants in terms of their class, structure, or meaning. Conceptual images are about being and essence, not doing and happening.

There are two primary types of conceptual processes: classificatory and analytical.

A classificatory process presents participants as entries in a taxonomy. It answers the question, "What kind of thing is this?" In these images, at least two elements are presented as belonging to the same category. Think of a chart showing different types of apples or a diagram of a family tree. The elements are separate but linked by a shared identity or superset. The layout often uses symmetry and hierarchy to establish these 'kind of' relationships.

An analytical process, on the other hand, relates participants in a part-whole structure. It answers the question, "What is this made of?" For an analytical process to occur, the image must contain two types of participants: a Carrier (the whole) and its Possessive Attributes (the parts). An anatomical diagram showing the organs in a body is a classic example, as is an exploded-view drawing of a machine.

Identifying these patterns isn't just a technical exercise. It reveals how visual texts distribute agency and power. This is where the visual grammar of Kress and van Leeuwen connects directly back to the concerns of Critical Discourse Analysis.

In a transactional narrative, the Actor is the agent of power. They are the one doing, acting, and causing change. The Goal, in contrast, is passive; it is the entity being acted upon. By consistently casting certain groups as Actors and others as Goals, visual media can reinforce social hierarchies and stereotypes.

For example, who is shown looking, and who is shown being looked at? The Reactor (the looker) often holds a position of knowledge or desire, while the Phenomenon (the looked-at) becomes an object of that gaze. This dynamic is central to discussions of visual representation in advertising, film, and art.

By learning to see these narrative and conceptual patterns, you can deconstruct the ideological work that images perform. You can move beyond what an image shows to analyse what it says—about who acts, who is acted upon, and how the world is organised.