Always Is a Claim, Not a Fact

Picture someone sitting in their car after a meeting that went badly. They are replaying it and saying some version of I always do this, nobody in that room respects me, and my boss obviously thinks I am out of my depth. Each sentence arrives with the full weight of a fact. Each one feels like an accurate summary of the situation, and that is exactly the problem. None of them is a report of what happened. They are compressions of what happened, and the compression has quietly thrown away the very details that would make the problem solvable.

That compression is the subject of this week's Mid-Week Mojo, and in NLP (Neuro-Linguistic Programming) the tool for reversing it is called the meta-model.

What language leaves out

Richard Bandler and John Grinder described the meta-model in 1975, in The Structure of Magic. It is a set of questions for recovering information that ordinary speech leaves out. Their observation was that people do three things to experience as they turn it into words. They delete: they leave out who, what and how. They generalize: one bad afternoon becomes always, and two unhelpful colleagues become nobody. They distort: a guess about what someone else is thinking gets stated as though it were something you know.

None of that is a defect. It is how a brain with limited bandwidth manages to talk at all, since a sentence that carried every detail would take an hour to say. The trouble starts when the summary is the only version we keep. We then make decisions, feel emotions and set expectations based on the compressed file rather than the original. The compressed file is usually bigger, darker and more permanent than what it replaced.

The questions that restore the details

The meta-model questions are simple. They are slightly irritating on purpose, because their job is to interrupt a sentence that wants to be accepted without inspection.

When you hear a generalization like always or never, the question is whether there has been a time it went differently. When you hear an unspecified nobody or they, the question is who specifically. When you hear a claim about someone else's mind, the question is how you know. When you hear I cannot, the question is what would happen if you did, or what specifically stops you. When you hear that something made you feel a certain way, the question is how specifically it did that.

Run those questions against the sentences from the car and something useful happens. I always do this becomes I lost my train of thought twice in one presentation after a short night. Nobody respects me becomes one person interrupted me and two people said nothing. My boss thinks I am out of my depth becomes my boss asked a hard question and I have assumed what it meant. The problem does not vanish when the details come back. It shrinks back to its real size, and a problem at its real size is usually one you can do something about.

What the research does and does not support

I want to be careful here. NLP attracts a lot of overclaiming, and I will not borrow the authority of science for it. NLP as a whole has a thin research base. A 2010 review by Tomasz Witkowski in the Polish Psychological Bulletin found little empirical support for its broader claims. The meta-model is best held as a practical language habit rather than a proven treatment, and it is not a substitute for therapy.

What gives the habit credibility is its close cousin. Cognitive restructuring is the thought-checking at the center of CBT (cognitive behavioral therapy). It targets many of the same moves, including overgeneralizing, mind reading and all-or-nothing thinking, and it has been tested far more rigorously. A 2012 review of meta-analyses led by Stefan Hofmann found strong support for CBT, particularly for anxiety disorders. That is evidence for the CBT version of the idea, not for NLP. The honest position is that the meta-model borrows a sound instinct without having been tested the same way.

If the always and never thoughts have become heavy and constant rather than occasional, real help from a professional is the right call. A set of questions on a blog is not.

For Leaders

Reports reach you already compressed. The client is furious. The whole team is burned out. Engineering never hits a date. Each of those may be true, and each one has lost the details you need to act well. Before you respond, ask who specifically, what exactly was said, and since when. Decisions made on the compressed version tend to solve a problem that is larger or different from the real one. The questions also teach your team that you want the original file, and that changes what they bring you next time.

For Individual Contributors

A great deal of the damage to your week is usually done by the sentences you say to yourself, and nobody else is going to ask them the meta-model questions. When you catch an always or a never in your own thinking about your work, write it down exactly as you thought it. Then ask it one question: whether there has been an exception, who specifically, or how you know. The sentence that has been running your afternoon rarely survives contact with its own details. What remains is usually a specific, fixable thing.

The rep

Today, catch one always, never, nobody or everyone in your own head. Write it down word for word. Ask it the matching question out loud, and write the answer underneath. See the compressed version and the restored version side by side.

Live your life uncensored.

Reference: Bandler, R., & Grinder, J. (1975). The Structure of Magic: A Book About Language and Therapy. Science and Behavior Books.

Reference: Hofmann, S. G., Asnaani, A., Vonk, I. J. J., Sawyer, A. T., & Fang, A. (2012). The efficacy of cognitive behavioral therapy: A review of meta-analyses. Cognitive Therapy and Research, 36(5), 427-440.

Reference: Witkowski, T. (2010). Thirty-five years of research on Neuro-Linguistic Programming. NLP research data base. State of the art or pseudoscientific decoration? Polish Psychological Bulletin, 41(2), 58-66.