In the early days of NLP's development, it was discovered that people use strategies for such things as making decisions or becoming convinced of something. These are not conscious strategies, but sequences of internal representations made up of visual, auditory, kinesthetic, gustatory and olfactory sensory components.

For example, in making a decision, one person might picture several options and say to himself, 'I like these two', then pick the one that feels best. (This would be a visual-constructed -> auditory-internal -> kinesthetic strategy.) Another person might prefer to first get a feeling for each of the options, picture how each might work out, then say to himself, "I like this one." (A kinesthetic -> visual-constructed -> auditory-internal strategy.)

Researchers Richard Bandler and Leslie Cameron-Bandler noticed that two people using the same strategy might arrive at very different results.

"For instance, two people might share a decision strategy with the structure: Vc -> Ki (deriving feelings from constructed images as a way to make a decision). One person, however, might report, 'I picture several options, and choose the one that feels right to me.' The other person, on the other hand, might complain, 'I picture several options, and then feel overwhelmed and confused by them.' [Both use the same Vc -> Ki strategy but get very different results.]

The notion of Meta Programs arose from attempting to discover what made the difference between such diverse responses. Because the general representational structure of the strategies was essentially the same, it was postulated that the differences came from patterns outside of, or 'meta to,' the strategy (or internal program); i.e., a 'Meta Program.'"

-- Dilts & DeLozier, Encyclopedia of Systemic Neuro-Linguistic Programming and NLP New Coding, 2000.

The initial set of Meta Program patterns was then identified and described by Leslie Cameron-Bandler in collaboration with David Gordon, Robert Dilts and Maribeth Meyers-Anderson.

Richard Bandler introduced the idea of metaprograms to NLP in the late 70s, as a way that people maintained coherence in their cognitive patterns.

Leslie Cameron-Bandler and others investigated further, using the Meta Model to identify a list of Meta Program patterns for use in therapy which eventually grew to around 60. One of her students, Rodger Bailey, simplified the model into 14 patterns for use in a business context – the LAB Profile (Language And Behaviour patterns) as set out in Shelle Rose Charvet’s excellent book Words That Change Minds.

The term ‘metaprogramming’ first appeared in John C Lilly’s book Programming and Metaprogramming in the Human Biocomputer (1968). Lilly presented the human nervous system as a biological computer, running ‘programs’ either hard-wired or learned. ‘Metaprogramming’ as Lilly describes it is changing the central control system so that we can learn more quickly and select more useful programs.

Tad James and Wyatt Woodsmall’s book Time Line Therapy and The Basis Of Personality presents a very similar simplified set of patterns, and links them to Jungian personality characteristics as used in the Myers-Briggs Type Indicator – not entirely convincingly in my view, but then I’m not a MBTI expert.