Audience design: the reader who shaped how you wrote
Allan Bell's 1984 proposal was that style is a response to the audience, the addressee most of all. That makes any single chat thread a narrow sample of how a person writes.
Audience design is Allan Bell’s account of why one person speaks or writes in several different ways. In his words, style is essentially a response to the audience, and speakers accommodate primarily to the person they are addressing. Third persons who are listening in shift the style too, by a smaller but regular amount.
It is a sociolinguistic idea from 1984, and it has a blunt consequence for anyone reading a chat export: the thread records one writer answering one particular reader, so what it shows is how that person writes to that reader.
What Bell proposed in 1984
Allan Bell published Language style as audience design in Language in Society, volume 13, issue 2, June 1984, pages 145 to 204. He was writing from Wellington and from the Department of Linguistic Science at the University of Reading.
The published abstract opens on a complaint. The style dimension of language variation, it says, has not been adequately explained in sociolinguistic theory. Bell’s proposal is that stylistic variation within one speaker derives from and mirrors the variation between speakers, and that the engine driving it is the audience. His sentence is that style is essentially speakers’ response to their audience, and that in audience design, speakers accommodate primarily to their addressee.
The idea that speakers converge on the person they are talking to was already the ground of accommodation theory, which the abstract lists among the article’s own keywords. What audience design adds is a ranking: the addressee first, and then everybody else, by how directly they are part of the exchange.
The addressee counts most, and the people listening still count
The abstract puts the rest of the audience in order. Third persons, auditors and overhearers, affect style to a lesser but regular degree. The effect does not vanish because someone is only listening; it shrinks in a predictable way.
Bell also folds in the things people usually put beside the audience. Nonaudience factors like topic and setting, the abstract says, derive their effect by association with addressee types. On that account you do not write differently about work because work is the topic; you write the way you write to the people you usually discuss work with.
The same account covers choosing between languages. Audience design, the abstract says, also accounts for bilingual or bidialectal code choices, which is the ground code switching in text messages stands on.
Referee design: writing towards someone who is not in the room
Not every shift is a reaction. Bell separates style shifts that are mainly responsive, caused by a change in the situation, from style used as an initiative, to redefine the situation that already exists. He calls the initiative kind referee design, and describes it as divergence from the addressee and towards an absent reference group. The abstract adds that referee design is especially prevalent in mass communication.
In a chat, referee design is the message written partly for a group that is not in the thread: the phrasing borrowed from a workplace, a subculture or a family register, aimed past the person reading it.
Why one chat is a narrow sample of a writer
If style is a response to the audience, then a single conversation holds a single setting of it. The person in that thread was writing to one addressee, with whatever auditors they believed were there. The style in the file is the style for that pairing.
This is a limit on reading, not a flaw in the writing. A person who is brisk in one thread and discursive in another is not inconsistent on Bell’s account; they are doing the ordinary thing, which is designing for a different reader. The useful question about an export is not what this person is like, but what this person is like with this reader.
Portraitor is built to the narrower question. It reads one named participant in one uploaded conversation and writes about that person, so what it describes is bounded by the same thing Bell describes: one writer, one audience, one thread. It has no other conversation to compare against, because it keeps none. What a chat analysis cannot tell you sets out the rest of that boundary, and reading your own messages back is the version of the question that stays inside it.
What an export carries, and what it leaves behind
An export is good evidence of the addressee and poor evidence of the audience. A two person thread names who was being written to. It does not record who was reading over a shoulder, who the messages were later forwarded to, or which absent group a particular turn of phrase was aimed at.
It is also thin on setting. Bell’s abstract makes topic and setting derive their force from the addressee types they are associated with, and an export preserves the topic while losing almost all of the setting: where each person was, what else was happening, how fast an answer was possible. Media richness theory is the older account of what a channel strips out, and why text messages get misunderstood covers what the reader supplies instead.
The practical reading of audience design, then, is a correction to a common mistake. A chat export is a sample of one register from one person, collected under one audience. It is real evidence, and it is narrow evidence, and those two things are both true at once. What your texts sound like to someone else works through the same gap from the reader’s side.
Frequently asked questions
Frequently asked questions
What is audience design?
Audience design is Allan Bell’s 1984 account of style variation, which holds that style is essentially a speaker’s response to their audience and that speakers accommodate primarily to their addressee.
Auditors and overhearers shift the style as well, by what the published abstract calls a lesser but regular degree.
Who came up with audience design?
Allan Bell, in Language style as audience design, published in Language in Society, volume 13, issue 2, June 1984, pages 145 to 204.
Is audience design the same as accommodation theory?
They are closely related and not identical. Bell’s abstract lists accommodation theory among the article’s keywords and says speakers accommodate primarily to their addressee, so accommodation is the mechanism at the centre of the model.
Audience design extends it by ranking auditors and overhearers below the addressee and by adding referee design, where a speaker diverges towards an absent reference group. Accommodation theory measured at scale covers the first half on its own.
What is referee design?
Referee design is Bell’s name for style used as an initiative rather than a response: divergence from the person being addressed and towards a reference group who is not present. His abstract says it is especially prevalent in mass communication.
Does a chat export show how someone writes to everyone?
No. On Bell’s account the style in a thread is designed for the person being addressed, so an export shows how one person writes to one reader rather than how they write generally.
That is a limit on what any reading of the file can claim, including a written analysis of it.
Read a thread as a sample of one register, aimed at one reader, and it tells you something true and narrow.
Your conversations, read the way you would read someone else’s.
Portraitor takes a chat export you already have, anonymises it in your browser, and returns a structured reading of how you come across. Identifying details are removed in your browser before anything is sent.
See how it works →