Annotation Scheme

At this moment no reliable procedure exists which can detect conventional metaphor in visuals. Some scholars have identified and analyzed conventional metaphors in images and movies while relying on their professional expertise and introspection, but a coherent and reliable method is still lacking. In order to offer a tentative classification of metaphor conventionality in visuals, we apply a rule of thumb procedure even though we are aware that this process still remains largely based on experts’ opinions.

Verbalizing a visual metaphor in a A-IS-B format is definitely a very difficult task for several reasons, one above all being: at which level shall I consider expressing the metaphor? Shall I take into account all the available information (including the time and setting in which the image appears)? Shall I stick to what is depicted? At which level of abstraction do source and target need to be expressed? We hereby rely on Steen’s model of metaphor, discussed in Sorm, Steen (in preparation), which distinguishes 3 dimensions of meaning: conceptualization, expression and communication. As for the communication dimension, for which Steen’s model of metaphor would distinguish a deliberate and a non-deliberate use of metaphor, we think that visual metaphors might always be deliberate, even though not always consciously produced as such. For this reason, at least for the time being, our annotation scheme takes into account only the two following dimensions: conceptualization (what do we understand?) and expression (what do you see?).

In principle, we believe that it might be possible to verbalize a visual metaphor at each of the three dimensions. The conceptualization of a visual metaphor is played on the chords of its perceived conventionality or novelty, while the expression of a metaphor is played on the chords of the perceived familiarity or creativity of the visual means that have been chosen to represent the metaphor’s domains. Moreover, also the visual realization of the metaphorical relation between the two domains is annotated on the dimension of the expression (see Forceville 1996, Phillips and McQuarrie 2004).

With VisMet 1.0 we hope to raise scientific awareness about this methodological issue. In particular, VisMetBaby’s aim is to facilitate the public sharing of comments and insights with the community of experts and non-experts that will help us produce a more reliable procedure in the future.

Before we start annotating

Step 1:

Before doing the annotation, take look at the information that is already available about the picture.

Step 2:

Begin the annotation by just looking at the entire image, including visual and verbal elements. What/who is being depicted here, what he is doing, where he is doing it, and so on. If the denotative meaning is ambiguous and allows more than one interpretation, then give alternative descriptions. We call this the expression level of the image. This level consists of: