Though it might seem artificial and perhaps forced to formalise informal arguments, this is a necessary step if we’re to come to some agreement on what counts as a good inference or recommendation. Evaluation requires formality, and here we’ll begin by developing a formal system to help us review the strength of the non-deductive arguments we put forth in investigations and recommendations.

Furthermore, this suggests that our evaluations of arguments must meet certain standards. As in any pursuit that requires judgements, application of standards might seem subjective or arbitrary to those new to the pursuit. In order to better understand the significance of the standards we’ll develop, and how best to apply them, you’ll need to practice. Fortunately, we run into informal inferences to the best explanation quite often, so we’ll have ample opportunity.

Let us begin by stating that, at a minimum, our inferences should be:

Minimum Standards for a Good Inference





These standards guide our assessment of an inference. We will get to the details of the standards later. For now, let’s observe that when we draw a conclusion, we can interrogate that conclusion by asking:

Interrogating for Minimum Standards

Is the conclusion consistent: does it make sense given the evidence at hand and not contradict any indispensable evidence?

Is the conclusion charitable: does it treat the evidence fairly?

Is the conclusion uncomplicated: does it tell a believable story without having to add unlikely details?

Is the conclusion comprehensive: does it address all the indispensable evidence?

These basic standards provide a foundation upon which we will build our formal system. First we must agree that, within our system, we will not accept conclusions (explanations or recommendations) that contradict the evidence at hand or deny obvious and agreeable characterisations of the circumstances. That is, we must agree that the best explanations available will be consistent with the evidence at hand.

For example, if we are investigating the cause of a fire, our conclusions cannot deny the fire itself. Rejection of evidence is unacceptable because our conclusions need to compete if we’re to evaluate one of them as best. If the conclusions address different sets of evidence, then they are not competitors, and this would undermine the very possibility of comparison. Comparing conclusions is the point of Inference to the Best Explanation.

Similarly, if I ask “Why won’t the car start?” and you tell me “because it’s not a car”, that is a correction of evidence, not an explanation. This answer does not compete with “it is out of fuel” or “there is an electrical problem”. Perhaps the investigation is better framed “Why won’t this machine start?”, but this would result from a separate evaluation of the assumptions I have made when framing my investigation, namely my assumption that this machine is a car and not a sculpture or a model.

Next, we must agree to do the best we can to represent evidence and resources with as little interpretation as possible. This is our principle of charity. To be charitable with respect to arguments is to make an effort to represent an argument as strongly as possible, despite personal opinions we might have about the content of the argument.

This is especially important in public debate, when misrepresentation of an opponent’s argument can weaken your contrary arguments. Misrepresentation of an opponent suggests to evaluators that you do not understand the position that you oppose. Of course, this is not to suggest that we be charitable out of selfish motives. Insofar as our task as critical thinkers is to advance the best explanations and recommendations possible, the only way to ensure we have considered our options thoroughly is to render them charitably.

Next, and derived from principles that underwrite scientific investigation, we aim for uncomplicated explanations and recommendations. From Aristotle to Newton to Kant to this text, we will prefer the least complicated among competing explanations. Note this doesn’t mean the least complicated will be the best. But in principle, we will prefer less complicated explanations over more complicated, all else being equal. (In a “Recommended Reading” box at the end of this section, we will include further resources that develop this topic.)

Finally, we aim to be comprehensive in our explanations. A best explanation should explain all the explainable and indispensable evidence available. An explanation that addresses all the available evidence should be preferred over one that doesn’t.

These are our principles, and to dispute them is to dispute the value of our system. Of course, they are not indisputable principles, and throughout history, thinkers have argued over the extent to which we should cleave to these and similar guidelines. Though Newton claimed that “Nature prefers simplicity”, it has not always been the simplest of explanations to win the day in the sciences. Nonetheless, these principles enjoy precedent in the history of ideas, and they have survived long bouts of intellectual upheaval. And so upon them, we shall build our system.

Last modified: Wednesday, 7 February 2018, 9:34 PM