Any account of image–language coupling must confront a familiar framework: the distinction between anchoring and relay, proposed by Roland Barthes.
In this account:
anchoring: text fixes the meaning of an image
relay: text and image contribute complementary parts of a message
This distinction has been widely adopted. It appears to capture something real about how images and language interact.
But it does not go far enough.
Anchoring and relay describe effects. They do not specify the structure of coupling.
To move forward, we must revisit—and rework—the distinction.
1. The Limits of Anchoring
Anchoring is typically understood as:
the reduction of ambiguity
the fixing of one meaning among many
This aligns with what we have already observed:
images are underdetermined
language constrains interpretation
But the notion of “fixing meaning” is too blunt.
Language does not simply select one meaning from a set already present in the image. It:
constructs the interpretive frame
specifies what counts as relevant
organises how elements are to be related
The image does not contain multiple meanings waiting to be chosen. It contains:
a field of potential construal.
Anchoring is therefore not selection, but specification under constraint.
2. The Limits of Relay
Relay is described as a division of labour:
part of the message is carried by the image
part by the text
Together, they produce a complete meaning.
This suggests:
symmetry
complementarity
additive combination
But as we have seen, the relation is not symmetrical.
Language:
specifies
organises
constrains
Images:
provide configuration
shape the field of possibility
The two do not contribute equivalent “parts” of meaning. They operate in different ways on different domains.
Relay, as usually formulated, obscures this difference.
3. From Functions to Relations
The problem with anchoring and relay is that they treat image–text relations as functions:
what does the text do?
what does the image do?
What is needed is an account of:
how the systems are related structurally.
This requires moving from:
- functional labelsto
relational analysis
4. Specification as Primary Operation
The central operation in image–language coupling is specification.
Language:
names elements
defines relations
situates the configuration
This does not merely “anchor” meaning. It:
constitutes the conditions under which meaning can be made.
Without specification:
the image remains underdetermined
interpretation remains unstable
With specification:
a particular construal becomes operative
5. Constraint and Range
At the same time, the image constrains language:
not all descriptions are equally plausible
not all interpretations can be sustained
This introduces the notion of range:
the image defines a range of possible construals
language selects and stabilises within that range
This is more precise than “relay”:
the systems do not divide meaning; they coordinate within constrained possibility.
6. Asymmetrical Coupling
The relation can now be restated:
image–language coupling is asymmetrical co-instantiation under mutual constraint.
co-instantiation: both systems are actualised together
asymmetry: language provides interpretability
mutual constraint: each limits the operation of the other
This formulation replaces:
anchoring (too static)
relay (too symmetrical)
with a structure that captures:
dependency
interaction
variability
7. Degrees of Specification
Different instances of coupling vary in how strongly language specifies the image:
strong specification
precise labels, technical description
minimal ambiguity
moderate specification
partial framing, suggestive captions
some openness remains
weak specification
minimal or absent text
interpretation remains broad
These are not different types of relation, but variations in the intensity of specification.
8. Reinterpreting Relay
What Barthes calls relay can now be reinterpreted more precisely.
In cases of so-called relay:
language may leave aspects of the image unspecified
the image may prompt further linguistic elaboration
But this is not a division of meaning. It is:
staged specification across systems.
Meaning is built through:
successive constraints
distributed operations
iterative construal
9. The Persistence of Misrecognition
Anchoring and relay persist because they align with intuition:
images seem ambiguous → text “fixes” them
images and text both contribute → they “share” meaning
But these intuitions:
conflate effect with structure
obscure dependency
flatten asymmetry
A more precise account must:
distinguish system types
specify operations
analyse coupling
10. A Fifth Specification
The argument can now be sharpened:
image–language relations are not instances of anchoring or relay, but structured couplings in which language specifies and constrains the interpretability of images within a bounded field of possibility.
Revisiting anchoring and relay does not require rejecting them outright. It requires:
repositioning them as partial descriptions
integrating them into a broader framework
replacing them where they obscure structure
What emerges is a more exact account:
images provide configuration
language provides specification
meaning arises through their coupling
The next step is to confront the strongest version of the opposing view:
that images, in themselves, constitute autonomous systems of meaning.
It is there that the stakes of this analysis become fully visible.
No comments:
Post a Comment