We find arguments in the wild frequently. News stories about aeroplane crashes draw conclusions from evidence. Juries recommend verdicts to courts. Advertisements try to persuade us that this product is the best for the job, from quenching thirst to driving across town.

In written form, we often identify that authors are making arguments by noticing keywords. Words like “therefore”, “so”, and “because” signal that the author is making some kind of an inference. Certainly there are occasions when these keywords, and others, do not signal inference, but they are good indicators. Let’s call them Indicator Words.

Some indicator words signal conclusions. Some indicator words signal reasons to think a conclusion is the best.

Indicator Words signal that an author might be giving reasons to think some explanation or recommendation is best. Common words include: therefore, so, because.

In general, when you see indicator words and, in the context, it’s clear the author is trying to persuade you of something, then there is likely some inference happening. These are informal arguments. When you identify inference is happening, then you can begin inspecting the inference, picking out reasons from conclusions, to see whether the conclusion is justifiable. The way to do this is to formalise the argument.

Note that you might start trying to formalise, then realise reasoning isn’t actually happening. In this case, either abandon the formalisation, or augment the prose to create a reasoned conclusion, as opposed to an opinionated one.

Opinions are distinguished by an absence of publicly assessable evidence. Without evidence, in the terms developed here, there can be no explanation. Given the inextricable connection between explanation and inference (as expressed in “Inference to the Best Explanation”), opinions are not and cannot be supported by argument. Of course, should you want to convert an opinion to an inference, you can add some evidence, plus some explanations. But beware: many opinions do not survive the scrutiny of well-structured inference.

For example, if I say “I really like your shoes”, that is my opinion; I have offered nothing more than a preference. I have given you no reason to think anything about those shoes, except that I like them. You might be tempted to think I could add: “I really like your shoes because I like red shoes”, and then I would be giving reasons to think something. Notice here we are still missing publicly accessible evidence, and that’s what we are looking for in Best Explanations.

In contrast, if I say “Those are the wrong shoes for walking around the zoo”, I can show you evidence of what happens to (other) people’s feet after walking all day in shoes like that. This evidence might give you reason to think that boots-made-for-walking would be superior to stilettos on a full day out.

Last modified: Wednesday, 7 February 2018, 9:23 PM