Final Words on Inference

We have been talking about the SPIRAL model since May, and this is the last installment on that. I have to thank you all for the emails on this topic. There is a lot to it, and you have clearly been paying attention judging from your excellent and insightful questions. I urge you to ask them IN THE BLOG, which will allow me to answer them there so that all can benefit. I will always answer emails, though and they are the best way to get an answer on a sensitive subject.

This time, we will address the two last types of inferences that we will cover: Analogy and Narrative. We have been covering the inferences from strongest (most desirable) to weakest, and so far have discussed inference from Example, from Cause, and from Correlation (also called Inference from Sign).

Analogies are a kind of comparison. We use them to compare a complex or unfamiliar concept that we want the listener to understand to a concept with which they are already familiar. The hope is that the listener will come to understand the unfamiliar concept by accepting that it is comparable to the familiar one. That acceptance of comparability is how we connect the evidence to the original claim, which is the purpose of an inference. Here is a quick example:

Our claim is – Nuclear reactors are very safe.

Our evidence is – the number of nuclear accidents since the beginning of the nuclear age compared with other kinds of accidents (like household).

Our Inference – A FAR larger number of accidents happen in American homes EACH MINUTE than the number of nuclear accidents that have happened since the beginning of the nuclear age.

Here we have used an analogy without actually stating something like “We can measure safety for nuclear reactors just as we measure household safety – by the number of accidents.”

Clearly this is a bad comparison – and a bad analogy. The potential impact of a SINGLE nuclear accident is many times higher than household accidents that occur in an entire year. One kind of “accident” does not compare with another kind of “accident”, even though they are both called “accidents”.

The test for a good analogy is simple “Do essential similarities outweigh essential differences?” Are there essential differences in the things being compared (in this case, household accidents and nuclear accidents)? When you read through the example, you may have “felt” there was something wrong with the argument but couldn’t put your finger on it. It is the essential difference between the two kinds of accidents that probably made you feel that. If you DIDN’T think anything was wrong with the argument, then you can see how you can buy into a bad argument by an analogy.

Analogies are weak because, at best, they are just comparisons of resemblances between things. You cannot “see” an analogy like you can see an example. You cannot test an analogy, like you can test a causal inference. You cannot measure an analogy, like you can measure a correlation. Still they are stronger than our last inference – Narratives.

A narrative is a story. It is designed to draw the listener in and give support to a claim by offering the listener with the opportunity to accept the story as adequate support for my claim if, and only if, the story is plausible.

Let’s say your claim is that “Hard work benefits those who engage in it” – and you offered as inference a narrative. We have all heard the fable of the Ant and the Grasshopper (if you haven’t, go here). It ends with the idea that “Just as the ant that works hard and doesn’t stop to play like the foolish grasshopper will be carried through lean times on the results of his hard work, you too will benefit by not resting while there is still work to be done.”

The story seems plausible – someone that works would likely be better off in hard times than someone that wouldn’t. The story is coherent. That is, all the pieces of it make sense with each other – nothing is inherently contradictory or counter-intuitive. The characterizations are consistent – the grasshopper doesn’t do anything strangely out of character, nor does the ant. And finally, the story resonates with us. It appeals to my sense of fairness that a hard-working person (or insect) should be rewarded for their effort.

And, these are the tests of a good narrative:

• Is the narrative plausible?

• Is the narrative coherent?

• Are characterizations consistent?

• Does the narrative have resonance?

You will find it very hard to convince someone of anything that is very important if all you have to offer is story. How about if I offered testimonials from a few hardworking people (or foolish grasshopper types) as examples? Or if I could demonstrate how hard work “causes” success? Or if I could show statistics indicating how hard workers do better in hard times? Or if I could compare the results of hard work and the results of leisure and draw an analogy with a relationship that you are very familiar with so that you could clearly understand it? Those inferences (again, from strongest to weakest, Example, Cause, Correlation, and Analogy) are all stronger.

If you are doing the talking, pick the strongest one you can. If you are doing the listening, test them as I have indicated. If you are refuting, use the strongest inference that you can AND ask why the other side hasn’t used stronger inferences to support THEIR side. Listeners will presume it is because there isn't any strong reason to support the other side.

Congratulations! This newsletter will be the last one on the SPIRAL model of critical discussions for a while. The topic for the NEXT series will be Persuasion. We will start on the fundamentals of persuasion next time!


Insist on great business results! Go to Pathfinder Communication