Ask Better Questions: How Not To Use The NLP Meta Model

My last posting about open and closed questions prompted this response from a reader, which got me thinking about how questions from the NLP Meta Model, or any kind of “challenging” question, have to be used with care if you don’t want to wreak havoc on your working relationships (he’s kindly given me permission to publish his question):

I was recently posed a question by an employee stating that something “Always” happened. I knew that it didn’t happen all the time and when I replied by returning the question “Always?” I was told by my peers that I was being aggressive and not adopting a coaching approach.

How should I, in your opinion, which I respect, should I have dealt with the question without closing the conversation and without sounding patronising?

What my correspondent did there was use one of the standard responses recommended by the NLP Meta Model for ‘universal quantifiers‘ such as ‘always’. The difficulty with using these standard responses is that so much depends on how you deliver them, and the context within which you deliver them.

That ‘always’ is an example of what is referred to in many books on NLP as a ‘Meta Model violation’. The very term ‘violation’ suggests something that is wrong, to which a ‘Meta Model challenge’ would be appropriate. Even if you don’t intend it, the bald use of a response such as ‘Always??’ may be received as aggressive or patronising, and consequently damage the relationship.

There’s an ‘early days of NLP’ story (which may or may not be apocryphal) about John Grinder teaching the Meta Model to his Linguistics class at the University of Santa Cruz, quite soon after he and Richard Bandler had come up with it. The students went out buzzing – they couldn’t wait to try out these amazingly precise questions and start uncovering the missing information and thought patterns that underlay the ‘surface structure’ of what people said.

The following week they shuffle back into the classroom looking like they’ve been through hell. They looked tired, scruffy, depressed; some of them had obviously slept in their clothes, others had been crying, and one or two had black eyes. Grinder asked them what had happened. “Well,” said one, “I tried out some Meta Model patterns on my wife, she was getting more and more pissed off, and after about half an hour she kicked me out of the house.” Every student had a similar tale of woe – “I had a stand-up fight with my best buddy”, “My family have disowned me” and so on.

That’s what the Meta Model questions will do if you use them without regard to maintaining rapport. Nobody likes being interrogated. Here’s how to use these very powerful questions to make your life easier rather than harder:

    1. Use the Meta Model as a guide to recognising particular language patterns that indicate missing information, over-generalisation, or ‘distorted thinking’. Have the appropriate response question flag up in your mind. But *choose whether or not you actually say it out loud*. Sometimes you may want to let the ‘violation’ go, or just bear it in mind as something to return to later.
    2. If you do decide to ‘challenge’ the ‘violation’ with a question, don’t ask the question in a challenging way. Use whatever ‘softeners’ you need to keep the relationship on track.

Sometimes it’s enough to just deliver the question in a softened, or humourous, tone of voice. Other times, you may want to ask the question in a different way, with some verbal ‘softeners’ around it. For example:

“So, I’m wondering if there’s ever been any times when there are exceptions – when something different happens?”

Or “I wonder if you’d mind me asking…” or “Maybe you could tell me if there’ve ever been any times…”

Or you could use a ‘reflecting back’ statement which will be taken as you checking your understanding, while at the same time posing the question in a non-threatening way: “So what I think you’re saying is, this happens every single time, and there have never been any exceptions?” which you could follow up with “Are there any circumstances in which it doesn’t happen?”

Again, you could soften these still further by adding “Let me just check my understanding…”. Or with the follow-up, you could say “Could there be any circumstances in which it wouldn’t happen?” which is hypothetical and unthreatening.

  1. Make sure that the “paralanguage” or non-verbal component of your conversation is right. I’ve already mentioned voice tone. How you stand or sit in relation to the other person is important too. You don’t want to be so close to them that you’re invading their personal space, getting “in their face” about the issue.Rather than facing them head-on (which is what the word ‘confrontation’ literally means) and across a desk (“keep your distance” or “I’m nervous of you so I will keep this barrier between us”), ideally you want to be standing beside them at a comfortable distance, so that you are ‘facing’ the issue together.

These kind of considerations would apply to using any kind of “challenging” question. Remember, your long-term relationship with the person is at least as important, if not more so, than retrieving any specific example of missing information or exposing a particular misconception in their thinking.

Meta model workbookIf you want to get really good at recognising Meta Model patterns and thinking of appropriate questions, you need to practice. A good study aid is Jonathan Altfeld’s Meta Model Exercise Workbook. It includes 12 sections (1 for each Meta Model pattern), and within each section, includes a Definition, Examples, Questions to ask, and 1 or more exercises to help you develop your skills. Find it here.

© 2010, Andy Smith. All rights reserved.


  1. Peter

    This is great! From someone who is DIY NLP Lite, sometimes I need these sort of things spelled out to me in more detail and this one has struck a chord. I too have suffered from questioning universal qualifiers in a too direct way. They do get p****d off eventually. I’ll try more softeners. Thanks.

  2. Stuart from systems thinking method

    Many thanks. As a consultant I constantly listen to the overgenerisations of clients who say things like "that never works" or "they always do that" I will respond by saying something like "come on, exactly how often does that happen?"

    Needless to say I lose a lot of rapport. Using a softener such as "may i ask a question just to clarify something?" is a real breakthrough in my thinking.

    Many Thanks.

  3. NLP


  4. Andy Smith

    Peter, you can get a long way with DIY NLP Lite as long as you pay attention to the responses you are getting and modify what you are doing in the light of the results you get!

    Stuart, thanks for your honesty. It's worth thinking of the verbal elements of a conversation as existing in the context of a wider system of non-verbal communication which carries information about the relationship you have with the other person – this is conducted by means of voice tone, body language, and also the type of questions you ask at a particular point in the relationship. If the other person doesn't think the question, or the way you asked it, is relationship-appropriate, this will cause perturbations in the relationship which could swamp the frail vessel of meaning that you want your words to be.

    I've just written a post about the role of non-verbal communication that you might be interested in.

    Asad, thanks for your interest!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.