Home › Communication theory
Problems with the encode–decode model
The code model explains how a signal travels, but not how people work out what others mean. Five problems stand out.
1. Meaning is not simply packed and unpacked
We say we “get an idea across” or “put thoughts into words”, as if words were containers. Michael Reddy (1979) called this the conduit metaphor and argued that it hides how much listeners contribute. The words It’s cold in here can be a statement, a request to close the window or a complaint, and the words alone do not settle which.
2. Listeners infer
H. P. Grice (1975) showed that speakers often mean more than they say, and that listeners recover the extra meaning by assuming the speaker is being cooperative. Asked “Is Sam a good cook?”, someone who answers “He has a lovely kitchen” has implied a doubtful answer without stating it. Sperber and Wilson’s relevance theory (1995) goes further: communication is ostensive–inferential. The code provides evidence, and inference does the rest.
3. Understanding depends on shared knowledge
Words such as that one, there or the usual mean nothing without common ground. Herbert Clark (1996) argued that speakers and listeners work together to build that common ground, and that they check it as they go, a process called grounding (Clark & Brennan, 1991).
4. Real conversation is not one-way
The model suggests a sender, then a receiver. In practice both people are active at once: listeners nod, say “mm-hm” and interrupt, and speakers adjust as they see reactions. Across ten languages, the typical gap between one speaker finishing and the next starting is only about a fifth of a second (Stivers et al., 2009), which means listeners are planning their reply while still listening.
5. “Noise” is more than interference
In engineering, noise is a physical distortion. In human communication, misunderstandings also come from differing assumptions, emotions, accents, cultural knowledge and word meanings. Nothing is wrong with the signal, yet the message still goes astray.
What the evidence suggests instead
| The code model assumes | Research on conversation shows |
|---|---|
| Words carry the meaning | Words give clues; listeners infer the rest |
| One person sends, one receives | Both people contribute all the time |
| Success means the message arrived intact | Success means mutual understanding was reached, often after repair |
| Errors come from noise | Errors also come from mismatched knowledge and assumptions |
Why this matters in therapy
If communication is a joint activity, the person with the difficulty is not the only one who can change things. Approaches such as supported conversation for adults with aphasia (Kagan, 1998) train communication partners to slow down, check understanding and use gestures, writing and drawing. In assessment it also helps to ask not only “was the message accurate?” but “what did the listener take it to mean?” For the mechanics of how turns fit together, see adjacency pairs.
Keep the model, drop the metaphor. The code model is still a good description of signals. It becomes a problem only when it is mistaken for a description of understanding.
Sources and further reading
- Reddy, M. J. (1979). The conduit metaphor. In A. Ortony (Ed.), Metaphor and Thought. Cambridge University Press.
- Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and Semantics 3: Speech Acts. Academic Press.
- Sperber, D., & Wilson, D. (1995). Relevance: Communication and Cognition (2nd ed.). Blackwell.
- Clark, H. H. (1996). Using Language. Cambridge University Press.
- Clark, H. H., & Brennan, S. E. (1991). Grounding in communication. In L. B. Resnick et al. (Eds.), Perspectives on Socially Shared Cognition. APA.
- Stivers, T., et al. (2009). Universals and cultural variation in turn-taking in conversation. PNAS, 106(26), 10587–10592.
- Kagan, A. (1998). Supported conversation for adults with aphasia. Aphasiology, 12(9), 816–830.