Philosophy of Writing – 16 – Dialogue I.

Posted on October 23, 2012 by


In linguistics it is generally considered to be the case that each participant in a conversation will assume a particular set of things about every other participant. To put it more simply: we expect things from the people we talk to.

A useful formulation of this “cooperative principle” of conversation is called “Grice’s conversational maxims”, after Paul Grice, a British language philosopher.

Grice said that in order to maximise the efficiency of a conversation, speakers will adhere to these rules and listeners will assume that speakers are adhering to them (there are exceptions, which we’ll get to).


Grice’s Conversational Maxims

Maxim of Quality
1. NO – saying that which you believe to be false.
2. NO – propositions that you believe to be unsupported by evidence.

Maxim of Quantity
1. Provide as much information as required by the current exchange.
2. Provide no more information than what is required by the current exchange.

Maxim of Relation
1. Be relevant. (Grice was unhappy with this rule.)

Maxim of Manner
1. NO – deliberately obscure speech choices.
2. NO – deliberate ambiguity.
3. Brevity.
4. Order.


I have found that, per my emotional life, Grice’s maxims are extremely important in conversation. For example: Boy Thing does not adhere to Quantity:2, Manner:1, or Manner:2 because he does not have standard concepts of them. He says whatever he damn well pleases, whether it be obviously relevant or no; he will say things to amuse himself rather than to contribute to a conversation. It. Drives. Me. Nuts. We have many arguments based on this. I only put two and two together long after I studied Grice and had gotten so frustrated trying to understand Boy Thing’s implications and what he meant by something or other that seemed irrelevant and why are you saying this to me and his reply because it amuses me and then I got it, finally: I was mistakenly assuming a maxim he didn’t even use. (Aside: Boy Thing’s response to Grice was “that’s stupid”.)

Real human beings are complicated. Short of direct brain-function capture before and after the corruption of memory, we’ll never have complete access to a person’s experience of the world, and therefore their whole personality. We assume Grice’s maxims because it’s the only real way to interpret conversations as information at all. Information must be assumed to have a certain level of quality for it to be meaningful.

For example: say a messenger is delivering something in person in the ancient world, when the fastest method for communication was still to bring a physical message from one place to another. If the messenger expends energy on a message to take it across many hundreds of km, if the message is contained in an expensive container, if the message is the only thing carried, then what can we assume? The message must either a) adhere to a certain level of quality, i.e. be genuine information that must be conveyed despite overall expense or b) contain any level of misleading information. So if it’s a meaningless jumble of letters, we can conclude that the messenger was sent there for some important reason and therefore his journey itself constitutes a kind of message.

When there is any kind of complex method of expense, it is usually safe to assume that the expense is producing deliberate communication and is therefore done for a reason. So, let’s say someone trips and falls on the street. This isn’t likely to be deliberate expense, even though people watching have received some very clear information. Now let’s say someone is muttering as he walks past you. Long before you dismiss his muttering as involuntary, you try to hear what he’s saying – you try to interpret it – because information is being conveyed at what seems to be deliberate expense. Note that speech is nearly always assumed to be deliberate, but that there are some forms of speech that aren’t: talking under the influence of something like sleep or hallucinogens, involuntary speech as a result of mental illness.

I’ve found another strange thing happening to me since the advent of headsets: I assume they’re talking to someone present, then I assume they’re suffering from a mental illness, and then I see the headset. This breaks the rule completely, because if Grice were complete as I quoted him I should assume the headset first. But that’s not the case, because Grice’s maxims operate under the strict thumb of probability. I grew up in a world where headsets were pretty rare – either they weren’t workable yet (when I was a child) or most people just used their mobiles – so I learned to first assume involuntary speech from someone distracted or from someone suffering a mental illness.

However, a writer is in the happy position of being intimately familiar with his characters’ personalities. So the writer can know exactly how and to what extent his characters will comply with Grice.

So to these rules, I might add my own meta-maxim, for fictional personalities only:


Maxim of Persona
1. Conformity to Grice’s maxims.
2. NO – conformity to Grice’s maxims more than appropriate to character.


So let’s say your context is a conversation between two cousins who have been friends since they were kids, and let’s say this is some of your dialogue:

“Well, Bob, because you’re my cousin, you should know that!” said Lacey.
“Sorry, Lacey.”
“It’s okay, Bob. Just don’t make that mistake again,” said Lacey.

This dialogue sucks because Bob and Lacey are comfortable with the idea that they’re cousins, and they both have meta-knowledge, i.e. they know that they’re both comfortable with the idea that they’re cousins. So Lacey, by saying “because you’re my cousin” does not fit with Grice’s maxim Quantity:2. From the background I gave there’s nothing to suggest that Lacey breaks Grice’s maxims due to her character – so her speech is unrealistic.

Note also that they use each others’ names over and over, which flagrantly breaks Manner:3.