Emotional reasoning

Emotional reasoning is taking a feeling as evidence about the world. I feel anxious, so this is dangerous. I feel guilty, so I must have done something wrong. I feel unwanted, so they must want me gone. I feel like a fraud, so presumably I am one.

The term comes from Aaron Beck's list of cognitive distortions, popularised by David Burns. It is worth separating from the version that circulates now, which is usually stated as your feelings are lying to you — a correction that is both wrong and useless. Feelings do not lie. They are highly accurate reports about your internal state. The error is in what they are treated as evidence for.

It rarely announces itself, because the conclusion arrives before the reasoning does.

The message has not been answered in four hours. You feel a drop, and by the evening you know something is wrong between you — not suspect, know. Nothing has been added to the situation except the feeling.

You feel unprepared before a presentation you have in fact prepared for, and conclude you will be found out. You feel awkward at a party and conclude you were awkward, which is a different claim requiring outside information you do not have. You feel like a bad parent on a bad day, and file it as a fact about the last decade.

Two forms deserve names because they behave differently. Forward: the feeling of dread is read as evidence that something bad will happen. Backward: the feeling of guilt or shame is read as evidence that something bad was done. The second is more convincing, because guilt feels so much like information about a past event, and it is the one that shows up hardest in the difference between shame and guilt.

It has been isolated in the lab. Arntz, Rauner and van den Hout gave people scripts in which objective danger information and a character's anxiety response were varied independently, and asked how dangerous the situation was. People with anxiety disorders inferred danger from the anxiety response even when the objective information said the situation was safe. Non-anxious controls mostly did not. They called it ex-consequentia reasoning — inference from the consequence back to the cause — and it is the cleanest demonstration of the mechanism there is.

Feelings are used as information in everybody. Schwarz and Clore's experiments found that people asked about their life satisfaction gave lower ratings on rainy days than sunny ones — unless their attention was first drawn to the weather, at which point the effect disappeared. Mood was being read as data about the life. This is the ordinary version of the same process, present in people with no anxiety at all, and it is why the distortion is so hard to spot from inside: using how you feel as a summary judgement is the normal shortcut, not an aberration.

The shortcut is usually efficient. That is the part the distortion literature under-sells. A feeling is a fast integration of an enormous amount of unarticulated information, and dismissing it wholesale throws away a genuinely useful signal. The failure case is narrow: it is when the feeling is being generated by something other than the situation — by tiredness, by a bad week, by what happened in an unrelated room an hour ago — and is then read as a verdict on the situation in front of you.

Here is the correction, and it fits in a line: a feeling is strong evidence about your state and weak evidence about the world.

"I feel anxious" is true, always, and requires no defending. "This is dangerous" is a separate claim, needing separate support. Most of the damage of emotional reasoning happens in the silent step between those two sentences, which is why the intervention is to make that step audible rather than to argue with the feeling.

This is also where the popular advice goes wrong in both directions. Trust your gut and don't listen to your feelings are the same mistake with opposite signs — both treat the feeling as a verdict to be accepted or overruled, rather than as one input with a known bias profile.

The bias profile is the useful part, and it is knowable:

• Feelings are least trustworthy as world-reports when you are tired, hungry, ill or between two and four in the morning. The entry on being tired but wired covers the specific unreliability of night-time conviction.
• They are least trustworthy about other people's motives, because that is exactly where you have the least data and the most filling-in. Hypervigilance supplies the readings; emotional reasoning turns them into conclusions.
• They are most trustworthy about long patterns. The feeling that something has been wrong for a year is a different kind of signal from the feeling that something is wrong this evening, and collapsing the two is the most expensive version of this error.

• Split the sentence. Say the feeling and the claim separately, out loud if possible. "I feel dismissed. I believe she is annoyed with me." The second one is now available for checking, which it was not while it was fused to the first.
• Ask what changed. If the situation is unchanged and the conclusion is new, the input was internal. That does not make it false; it makes it unproven.
• Give the claim a test. What would you see if it were true, that you would not see otherwise? Most emotionally reasoned conclusions have no answer to this, and noticing that is often enough.
• Postpone the verdict, not the feeling. The feeling is allowed to stay. The conclusion can wait until morning, and a surprising number of them do not survive the wait.
• Raise the resolution. Emotional granularity is the upstream fix: "bad" converts into a verdict about your whole life far more readily than "let down by one person about one thing" does.
• Notice the loop version. When the conclusion keeps being re-derived from the same feeling every night, this has stopped being a reasoning error and become rumination, which needs the other entry's answers.

The Overthinking test reads how much of your attention goes into re-deriving conclusions, and the Anxiety Check reads the worry that most often supplies the feeling in the first place.

The scope note: this is a description of an ordinary thinking shortcut, not a treatment, and cognitive therapy is more than the list above — it is a structured course of work with a clinician, not a paragraph. If the conclusions are about your own worthlessness, and they hold at all times of day rather than arriving with a mood, the useful move is not better reasoning. That is depression's characteristic signature, it is treatable, and it belongs with a professional rather than with a worksheet.