I don't see any other way of interpreting your claim that 'the way to get the most value out of flawed people is to find the best version of their argument and then argue with that.'

You give your argument, they give theirs. Then you amend their argument for them, and argue against that, ignoring your interlocutor.

If you are not interested in accurately understanding the person you're talking to, and instead impose what you take to be the strongest interpretation of their argument on them, then you are for all intents and purposes arguing not against them, but yourself.

> Then you amend their argument for them, and argue against that, ignoring your interlocutor.

The problem is that language is ambiguous and easily misunderstood. You interlocutor also has context which you don't know about.

Being charitable means recognizing these facts and so reading their words to mean the strongest possible interpretation. If they intended some weaker argument, it doesn't matter, because dismantling the stronger version already handles the weaker interpretation.

I suppose the problem in all this is that the dictum that you should 'criticise the best possible version of your interlocutor's argument' does not specify how to identify the argument in question in the first place.

Take the following example. Suppose you claim that the meat industry is bad, advance three premises, and infer from them the conclusion that it is indeed bad. Now, if I use the dictum 'criticise the strongest version of your interlocutor's' argument, that could mean many things:

  1. I might think that all the premises were unconvincing and argue against entirely different premises in favour of the same claim.
  2. I might offer new premises in addition or substitution for the premises provided, or a different conclusion.
  3. I might interpret the specific details of the premises or conclusion differently.

One could plausibly say that all of these strategies offer a different version of the 'same' argument, but the first would have almost nothing to do with it, and the second would be majorly different. Also, note that if I switch around the premises, and argue against that amended version of their argument, it does not at all follow that 'dismantling the stronger version already handles the weaker interpretation'. Defeating a strong argument does not falsify an entirely independent argument, even if it is weaker.

I personally would say that this conversation is based on an unhelpful conflation. We should both strive to understand one another as accurately as possible, and to move the conversation in as constructive a direction as possible. They need not be in tension. The problem arises if I take the latter - the desire to make the conversation constructive - as an interpretive principle.

I would say that none of your cases covers the intent behind the dictum. You've already moved past the "ambiguous language" and right into assuming that the argument was clearly conveyed and understood, and so you must now endeavour to evaluate the plausibility of the premises, the inferences, etc.

The dictum applies prior to this step, where the premises and the argument may be phrased in a somewhat ambiguous way. When interpreting the language used to describe the premises, take the most challenging interpretation rather than the weakest/most easily attacked interpretation. When evaluation the inferences, once again take the most challenging interpretation.

For instance, a phrase may admit two interpretations, a valid or invalid inference. You should always assume your interlocutor intended the valid inference.

If you find your interlocutor's position ambiguous in some major way, then you should, firstly, tell them, and secondly, ask them to explain. Otherwise confusion may ensue as to what exactly is being argued for and against. If the ambiguity is incidental, then the principle of charity is not really doing much. I take it that your example - about a phrase admitting a valid and invalid inference - is of this kind. Though I do agree that we shouldn't jump upon those kinds of slippages.

If you read carefully and allow the conversation to unfold, while prompting your interlocutor to explain what remains unclear to you, then generally you can get a fairly accurate impression of their views. More often than not, the problem is not ambiguity, but an unwillingness to listen, and engage thoughtfully with others, through a genuine dialogue.