Article Categories
[ Show ] All [ Hide ]
Clean Language
Article Selections
[ Show ] All [ Hide ]

These notes were first presented at The Developing Group 2 October 2008

Vectoring and Systemic Outcome Orientation

Penny Tompkins and James Lawley 

Whatever happens during a session, excellent facilitators and therapists always seem to know where to go next. They are also able to pursue a line of questioning and to navigate elegantly through the client’s information.  To find out how they do this we undertook a modelling project. Our exemplars were David Grove, Steve De Shazer, Robert Dilts, Steve Andreas (and ourselves).   

We were particularly interested in contexts where the facilitator worked in a bottom-up, systemic way, i.e. instead of starting with a predetermined end point or technique, the outcome emerged out of the interaction during the process.

While this paper focusses on how vectoring works in Symbolic Modelling and other derivatives of David Grove’s work, vectoring is applicable to many situations including group work, training, chairing a meeting, interviewing, sales, research, etc.

This paper is in five parts:
  1. The Model
  2. Common Vectors used in Symbolic Modelling
  3. The Structure of a Vector
  4. Sample Vectors
  5. Some skills required to facilitate using Systemic Outcome Orientation

The diagram above represents how a facilitator-therapist can use vectoring. Since it is a process which involves “circular chains of causation” (Bateson), and since circles do not have a ‘beginning’ or an ‘end’, you can arbitrarily start anywhere and follow the flow around the system.

For convenience, in the description below, we will start with the facilitator’s initial state, #2:

First Round

#2 The facilitator starts with a blank or empty model of the client’s model (map) of their world.
#3 That means starting with no idea of the client’s desired outcome or any predetermined ‘ideal’ state (note: this includes goals of many types of facilitation such as “integrated”, “congruent”, “mature”, “in their body”, “aligned”, etc.).
#4 The facilitator chooses a desired process outcome. Rather than this being a specific end point the aim is for the client’s attention to move in a direction, i.e. on a ‘vector’ (see below). An example of a common starting vector is to aim: ‘for the client to identify a desired outcome (if they have one)'.
#5 The facilitator asks a question which has a high likelihood of meeting the aim in #4, e.g. they ask “And what would you like to have happen?”.
#6a The client responds internally (based on their model of the world).
#6b The client responds externally (verbally and nonverbally).

Second and Subsequent Rounds

#1 The facilitator notices some of the client’s verbal and nonverbal responses (#6b). Penny for example is "paying attention to what the client is paying attention to" from their perspective.
#2 The facilitator updates their model of the client’s model (map) of the world based on those responses.
#3 The facilitator notes the client’s desired outcome statement — or the lack of one — and makes it a reference point for all future decisions.
#4 The facilitator chooses a desired direction, a vector, for the next part of the process. If the facilitator is using the Problem-Remedy-Outcome model as a guide then:
  • If #6b does not include a desired outcome statement, the facilitator continues on the same vector as #4 in the first round.
  • If #6b does include a desired outcome statement, the facilitator switches vector, e.g. ‘for the client to develop their desired outcome statement into an embodied (symbolic) perception’.
#5 The facilitator asks a question (or does some other behaviour) which has a high likelihood of meeting the chosen vector.
#6a The client responds internally (based on their model of the world).
#6b The client responds externally (verbally and nonverbally).
And so on. Now the process repeats iteratively, i.e. each round builds on the previous state of the system.

Over Time

Let’s look at how the process works over time:

The facilitator’s model of the client’s model, #2, is being updated each round and therefore takes into account the entire history of their observations (#1’s). This means #2 has to be a dynamic model that incluedes a sense of the progression (directionality) of the client’s system.

During the course of a session a client will often  identify multiple desired outcomes and/or multi-part outcomes. These will often be revised by the client during the session. Thus #3 over time contains the history of all the client’s stated desired outcomes and a sense of the progression (directionality) of those outcomes, i.e. #3 has to be a Dynamic Reference Point.

The client’s desired content outcome, #3 may start off as a single verbal statement but over time will be elaborated into a richer description of: external behaviour; VAK internal states, criteria for achievement, strategies for maintaining the desired outcome; metaphoric/symbolic representations; etc.


The facilitator is always on some vector, if not chosen consciously, then chosen unconsciously. David Grove liked to quote his mentor, the sociologist Bill Rawlins, "whatever you are doing you are always up to something".

At #1 the facilitator notices what just happened in relation to the question just asked, #5. This creates a key feedback loop. There is information in whatever is said and done, and their is information in how that relates to the previous question. For example, what does a client have to do internally to answer a request for a desire with a statement of a problem?

We presuppose that a facilitator asks a question at #5 which is likely to head in the direction of their own desired vector #4. This isn't always the case, a facilitator might have an intention to be on a vector to model the client’s desired outcome – and actually be doing something else, like modelling their problem. However, if the facilitator’s intention is to model the problem then that is a vector in its own right.  It is important to note: vectoring is not just about going for outcomes.

Skills Required

To use vectoring a facilitator needs to develop certain skills, such as the ability to:

#5 Know the function of each question they ask and be able to deliver it with the appropriate syntax, vocal qualities and nonverbals.
#6 Know what a client is likely to do internally to make sense of the question. In other words, when/where in the client’s perceptual landscape does the question likely direct their attention?
#1 Be able to distinguish a client desired-outcome statement from every other kind of information presented e.g. using our REPROCess model this might be: Resources, Explanations (including meta-comments and ‘facts’), Problems, Remedies, Changes.
#2 Separate their model of the client’s information from their personal model of the world (including their model of therapy).
#3 Hold the client’s current desired outcome(s) in mind throughout the session.
And, be able to track the history and directionality of the client’s desired outcome statements throughout and across sessions.
#4 Have knowledge of, and the skill to apply a range of common vectors, and be able to devise a vector in-the-moment based on the logic in the client's information.

(See below for a list of common vectors used in Symbolic Modelling.)

At any moment there are three kinds of outcome:

The client's desired content outcome. [Note 1]
The facilitator’s model of the client’s desired outcome — the facilitator’s Dynamic Reference Point.
The facilitator’s desired process outcome — which is in service of #6.

This can be complicated if a client also has a process/means desired outcome, e.g. "I want to achieve X by means Y." Commonly a client will choose a means that either won’t get them to their desired outcome or reduces their options. This is when the Solution is a Problem (Paul Watzlawick).
The effectiveness of the vectoring will depend on:
  • The isomorphism (structural similarity) of #3 and #6.
  • The relevance of #4 to #3.
At times the process requires subtle modification, when:
  • The client cannot state a desired outcome.
  • The client is running a Self-Delusion, -Deception, -Denial pattern.
  • The facilitator decides the client's desired outcome is seriously unecological.
There are a number of other distinctions that are vital to keep in mind when working in an outcome orientated way:
  • The difference between a desired Outcome and a proposed Remedy (attempt to solve a Problem)
  • The difference between a desired outcome and an actual outcome? [Timeframe]
  • Who 'owns' the outcome and who is it for? [Perceptual Position]
  • Is it a content or process outcome? [Level]
Let’s look more closely at the ownership and the content/process distinctions.

A desired content outcome will have an owner and be for someone. There are four options:
  • Client's desired outcome for them self or someone else in their life (needed for outcome-based therapy/coaching.)
  • Client's desired outcome for the facilitator (e.g. The client plays 'please the facilitator').
  • Facilitator's desired content outcome for the client (needed for therapies that have a predetermined outcome; and for administrative aspects, e.g. “I want the client to pay me”).
  • Facilitator's desired content outcome for them self. (Needed in product/exemplar modelling where there is no intention for change; or for exploitation where the intention is to benefit the facilitator at the expense of the client.)
A desired process outcome will be set by the facilitator (except where a sophisticated client wants to decide the process as well as the content - in which case the facilitator needs to work at a meta desired process outcome level). Different styles of facilitation will be based on the dominant timeframe of the desired process outcome:
  • In the moment
  • For the next several minutes (= a vector)
  • For as long as it takes to do a technique (consisting of a number of vectors)
  • The session
  • Beyond the session
Expert facilitators are simultaneously aware of all these timeframes, although they may or may not use formal techniques.


A vector is a facilitator’s desired process outcome for the client.  It involves movement in a direction. Vectors are usually general enough to allow for several means of moving along a vector.  While there are common, off-the-peg vectors which are used over and over, some vectors are 'bespoke' in that they are decided in the moment and are designed to fit a particular client's circumstances. As a rule of thumb, vectors are usually about 3-10 questions in length.

The purpose of using the metaphor ‘vector’ is to emphasise that the process is going or heading in a direction. In a bottom-up modelling process the ‘end’ of a vector is rarely reached. More likely a new vector is chosen part-way through. There is a strong analogy with a sailing boat tacking (see diagram).

It is vital that the distinction between a client's desired content outcome and the facilitator's vector (desired process outcome) is clear to the facilitator. Should this be forgotten there is an increased likelihood that the facilitator’s desired content outcome becomes involved, and the process  becomes less and less clean.

A key concept is ‘nested or simultaneous vectors’.  For example, in a recent demonstration, Penny's initial vector was "For the client to identify a desired outcome" (vector A).  Since this wasn't immediately forthcoming, Penny kept this vector open while heading off on another vector which was "For the client to develop a perceptual landscape for their current state" (vector B). As Penny facilitated the client to locate a few symbols a current-state Landscape began to take shape.  When, at her tenth question, Penny again asked ”And what would you like to have happen?”, the client answered with a desired outcome statement. This ended vector A. Vector B was left open and temporarily set aside in pursuit of a new vector "For the client to develop their desired outcome perception" (vector C). This continued for half a dozen questions until a problem with the desired outcome became apparent, which set Penny off on a new vector D ...


There is a similarity between a vector and a TOTE. The T.O.T.E. (Test, Operate, Test, Exit) model is an iterative problem solving strategy based on feedback loops devsied by George A. Miller, Eugene Galanter, and Karl Pribram (see Plans and the Structure of Behavior, 1960). If you are familiar with TOTEs and are able to apply them in the moment during a client session you probably won’t need the notion of a vector. However, if you’ve never quite been able to apply TOTEs, the vector metaphor might be more to your liking. One advantage of the vector metaphor is that it is more embodied than the TOTE — imagine sailing boats making progress by tacking rather than abstract computer algorithms.

We’ve always found it strange that although a Goal is part of the model it is not included in the acronym TOTE.  By contrast,  a vector by definition includes a direction. Another distinction between the two is that a vector is about heading in a direction, whereas a TOTE is about achieving an end-state (from which you Exit).  In the ever-changing process called therapy, if you are not using a technique, getting to the ‘end’ of a vector is an exception rather than the rule.
To conclude

At every moment vectoring requires the facilitator to hold in mind both a client-content and a facilitator-process desired outcome. Because the client’s desired content outcome can change over time, and because except in the simplest of cases the client can't go straight from where they are to their goal, the facilitator needs to regard the client’s desired outcomes as a series of dynamic reference points.

In a top-down, technique-based approach the facilitator has an idea of where the whole process is going and their job is to guide the client towards that.

In a bottom-up vectoring approach the end result is not known until the client gets there. Therefore the facilitator needs to continually be prepared to modify the process direction in light of each new piece of information. In setting a direction in the moment the process heads along a vector for a short time before heading off on the next vector. The overall direction of the session thus becomes the sum of all of the vectors and the length of time each vector is maintained.

Lastly, in the above we have focused on desired outcome orientation. Another important component is actual outcome orientation. As a rule the focus of the orientation will over time alternate between desired and actual outcomes.

SOURCE MATERIAL of our exemplars:

Steve Andreas, a video of Steve using The Forgiveness Pattern (CD set).

Robert Dilts, the transcripts and videos of the Northern School of NLP workshop (2006) examples available in Modelling Robert Dilts Modelling.

Steve De Shazer, the transcripts in Words Were Originally Magic (1994).

David Grove, many many hours of observation and the all the transcripts, audio and video tapes listed in his Bibliography of Publications.

With thanks to Phil Swallow for coming up with the term 'vector' and for numerous other contributions to this model. Contributions also gratefully received from Marian Way, Matthew Dodwell, Wendy Sullivan and Judy Rees.


Penny Tompkins & James Lawley
Penny and James are supervising neurolinguistic psychotherapists – registered with the United Kingdom Council for Psychotherapy since 1993 – coaches in business, certified NLP trainers, and founders of The Developing Company.

They have provided consultancy to organisations as diverse as GlaxoSmithKline, Yale University Child Study Center, NASA Goddard Space Center and the Findhorn Spiritual Community in Northern Scotland.

Their book,
Metaphors in Mind
was the first comprehensive guide to Symbolic Modelling using the Clean Language of David Grove. An annotated training DVD, A Strange and Strong Sensation demonstrates their work in a live session. They have published over 200 articles and blogs freely available on their website:
 »  Home  »  Modelling  »  How to Model  »  Vectoring and Systemic Outcome Orientation
 »  Home  »  The Developing Group  »  Vectoring and Systemic Outcome Orientation
Article Options

et le
clean language

James Lawley and
Penny Tompkins

16-18 June 2017

more info
in French

view all featured events