NLP

NLP is said to be recursive. What does that mean?

From a conversation on NLP connections:

innovating new high-level models in NLP? … is expected to be out of reach for most.

Richard also encourages people to go out and invent their own stuff. I don’t really think NLP is a field, the way, say, optics is a field. NLP is a particular approach to psychology, but to become self-sustaining, it needs enough organizational, research, and development infrastructure to outlast its creators. That infrastructure just isn’t there.

Bizarrely, EMDR might end up lasting longer, more widely used and accepted, than NLP, even though Grinder claims EMDR’s Francine Shapiro basically took one tiny NLP concept, relabeled it, and went out and sold it (successfully) to the established psychological machine.

So as an open question, what specific parts of computer science … to build recursive and generative models

Well, for one, go study recursive functions in computer science. If you work your way through the book Structure and Interpretation of Computer Programs by Sussman and Abelson, you’ll know recursion like the back of your hand. (You actually have to do the problem sets, though. Understanding it is very different from being able to do it. The book is now freely available on the web.)

Recursion is the programming equivalent of mathematical induction, by the way. It’s hard to understand without learning it deliberately. And it’s hard to use without a learning curve. Many people throw the word around with no understanding or concept of what it means. I used to TA a course in computer science taught using the programming language LISP. LISP is notable, among other reasons, because it’s designed to find recursive solutions to problems. Some people “got it,” but many people didn’t, even very, very smart people.

Recursion is defining a function in terms of itself. Consider the Fibonacci series: 1, 1, 2, 3, 5, 8, … Each term is made by summing the previous two terms. If you were to try to write a normal math function that would generate the sequence, it would be very long and complicated.

But you can express it recursively very simple:

The Nth fibonacci number = The N-1th fibonacci number + the n-2th fibonacci number
or in math terms: Fib(N) = Fib(N-1) + Fib(N-2)

If you think about this, it works, except somewhere you need to specify two consecutive Fib numbers so all future ones can be computed. The final recursive definition looks like this:

Fib(N) = Fib(N-1) + Fib(N-2)
Fib(1) = 1
Fib(2) = 1

So now we know:
Fib(3) = Fib(2) + Fib (1) = 1 + 1 = 2
Fib(4) = Fib(3) + Fib(2) = 2 + 1 = 3
etc.

In some sense, the later Fib() functions are built up of earlier Fib() functions.

In LISP, recursion is sometimes used to build self-modifying programs. Since the program is essentially defined in terms of its earlier self, it’s recursive.

How does recursion apply in NLP?

One place recursion applies in NLP is in strategies. A non-recursive strategy is attached to a specific time and place. For example, “Go into a bar. See attractive person. Feel confident. Walk up and offer to buy a drink.” There’s no recursion in that.

A recursive strategy in some way refers to itself or builds a stronger version of itself. For example, “Go into a bar. Remember what you did last time and generate a dozen new possiblities for how to behave to meet someone. Go do that. Afterwards, future pace that behavior if applicable to a dozen new contexts.” That’s recursive because the behavior in any one time is built on past runnings of the strategy. The strategy is also self-modifying and self-improving.

In Richard’s trainer’s training, he teaches some specific skills. A big piece of the strategy he installs, however, is essentially, “When in a troublesome situation you’ve never been in before, enter a resourceful state and invent a new strategy on the fly.” That’s recursive because it’s a strategy-modifying strategy.

Make sense?

When Bandler, Grinder, Dilts, etc. first formalized NLP, they certainly did it using the terminology of mathematical models, particularly calculus.(1).

The early NLP literature (Structure of Magic, NLP Vol I) talks about “4-tuples” (or 5-tuples, if you go back far enough), and “operations” that can be done to 4-tuples.

This was an early attempt to make NLP a calculus. Math-phobics, suspend that phobia. We aren’t talking mathematical calculus, but a general calculus. That just means they defined distinctions to pay attention to, operations on those distinctions, and rules telling how they all went together and what produced what results.

In Math, distinctions called “numbers” include 4, 5, and 9. Our operators include something called “addition.” The rules of math say when you combine 4 and 5 using addition, you get 9.

In NLP, we have distinctions called “4-tuples” with a specific set of internal/external sight, sound, small, taste, etc. We have an operators, “set anchor” and “fire archor.” Given two different 4-tuples, we can anchor both. When we combine them using the operator “fire off anchors,” we get a new 4-tuple with elements of the original two.

So NLP, itself, is a model. It has distinctions like 4-tuples, physiological state, internal images, auditory voices, submodalities, the unconscious mind, etc. It has operators like anchoring, shifting submodalities, etc. It also has rules for how those combine: if someone has Friends coded in one set of submodalities and Acquaintances in another, they can make an Acquaintance a friend by shifting the set of submodalities.

When they were developing NLP, existing therapies weren’t this rigorous. They didn’t have well-defined distinctions or operators, and had no real idea why or when their stuff worked. NLP was (and to some extent, still is) novel in that it attempted to be as rigorous as a mathematical model.

(To this day, the DSM-IV, the traditional therapeutic Diagnostic and Statistical Manual, has many descriptions that are too vague to be used as good, rigorous distinctions.)

Producing models with NLP

NLP can also be used to produce models. You can use the NLP distinctions to build a model of a skill. The famous spelling strategy, for example, uses the NLP distinctions to produce a model of how some good spellers spell: they create a mental image of a word properly spelled and anchor it to the sound of the word. That’s a super-simple model, but it reflects how the process operates.

B&G originally hoped people would go out and use NLP to produce models of how people did all sorts of skills. In practice, this hasn’t happened. I’m not sure why, but I suspect that successful modeling is a specialized skill that isn’t terribly useful in daily life, so few people get good at it.

Models by themselves aren’t very useful. Their usefulness comes from applying them. You can develop one model and then spend a lifetime applying it. The paradox is that those who love building models rarely enjoy applying them once the model seems to work. And those who like application are rarely good at building them.

Some professions are pure model-building professions. Linguistics, mathematics, physics, computer programming, academic research, and some forms of management consulting (e.g. business process re-engineering) are all model-building professions. Look closely at that list and you’ll find that those professions sort by task/system, not by people. They may be good at building models, but those practitioners rarely spend their time developing fine distinctions about people. Rather, they model things and systems.

So is NLP a model? Yes.
Can NLP be used to build models? Certainly.
Is NLP used to build models? By a few people, but rarely.
Is NLP necessarily model-building? Not at all. You can be highly skilled at using NLP for therapeutic interventions and not do any model-building.

Addendum from a conversation on 18 Aug 2007

A key piece of modeling is choosing the distinctions your model will have. Physics, for example, uses the distinctions “FORCE,” “MASS,” and “ACCELERATION.” Newton is very famous for finding and proving the relationship FORCE = MASS * ACCELERATION (F=MA).

What is often (always?) overlooked is that the very choice of Force, Mass, and Acceleration to measure is, itself, genius. If he had chosen WEIGHT, DENSITY, and SPEED, he likely would have found no relationship.

Bandler’s greatest genius, in my mind, is that he simply slices up reality a bit differently from the rest of us when watching people. He creates new models not because he has great skill in modeling (though that helps), but because he can slice up his observations in ways no one has ever done before. His ability to articulate what he does is relatively rare, and lets him teach portions of it.

For example, he noticed voice tone and tempo when talking to Erickson. Others simply hadn’t noticed it before. Is the genius in noticing that Milton would embed commands through his tone or tempo (relatively easy to hear, once you know you’re listening for tone and tempo changes)? Or is the genius noticing that tone/tempo might be relevant in the first place.

Maybe it’s both.

That’s why I find Bandler irreplaceable in many ways. He perceives differently, and that is a powerful piece of his modeling.

I came across this when modeling software engineers long ago. After many frustrating hours trying to figure out how one superb programmer broke down his solutions into code, he simply shouted (words translated to a metaphor for the non-programmers), “Stever, you just don’t get it. HAMMERS aren’t used to pound nails, they’re just a way to provide bracing while you use the door frame to pound the nail.”

His definition was radically different from how everyone I knew thought about hammers. But with his definition (his way of slicing up the world), many previously hard problems suddenly became simple. Ditto for the entire concept of recursion, by the way. Many extremely hard problems, when expressed recursively, can become absurdly simple.

Modeling = the distinctions you make AND the relationships you find between them.

The magic resides in both halves of the definition.

(1) For those of you who have taken group theory, think “Rings” and “fields.”back

What’s the difference between the two?

My impression is that unconscious assimilation is the natural ability we have to mirror others to the point where we adopt their cognitive skills. It’s the way we learn language as an infant, and it extends well into adult behavior (ever notice that people follow their leaders’ actions, not words?) For some fascinating reading on what might be the mechanism behind unconscious assimilation, Google “mirror neuron system.”

Assimilation isn’t installation, at least not as Bandler uses the word.

Installation, as Bandler uses the word, is very precise, and very different. Bandler leads audiences through sequences of unconscious representations and emotional states over and over and over until those sequences get learned through repetition. Then he anchors them to the desired context.

For instance, I once saw Bandler up on stage (ever seen his Orlando videos? Where he’s sitting there with a huge banner saying “NLP Seminars Group” above his head?) as he was doing an installation. He was talking about how he remembers a given person’s meta-programs and issues. He does it, he says, by visualizing a list of their criteria, etc. above their head. So looking at someone, he sees this little checklist above their head, telling what work needs to be done. Of course, journalists just use a little book and jot notes. It’s an archetypical American image: the 1950s journalist in dapper suit, gripping a notepad and scribbling with a pencil. Sometimes, they even have a pencil stuck behind their ear, right beneath a hat with a big sign in the hatband that says “PRESS.” Very film noir.

If you followed along that last paragraph, three times you likely made pictures of people with large, readable words over their heads. If we told those stories a dozen times, your unconscious mind would start to “get” the general strategy: make pictures of big, legible words over people’s heads.

If I anchored it correctly, later I can help you link it to the right context. I might ask, “If you’d like to learn a cool way of remembering people’s names, stop for a minute and imagine you just met someone on the street…” As I see you start to enter your name memory strategy, I fire the anchor and Bazoom! the strategy gets linked. I would reinforce it with some hypnotic language, test the work, of course, and probably tell your conscious mind some other story to distract you. The installation part is the rehearsing and anchoring of the strategy and associated feeling states.

So in my understanding:

• Assimilation uses the mirror neuron system to “get it” unconsciously.
• Installation is a deliberate leading of someone through representations below conscious awareness to help them develop a cognitive skill.