NLP holds it as crucial in human cognitive processing to recognize that the subjective character of experience is strongly tied into, and influenced by, how memories and perceptions are processed within each sensory representation in the mind. It considers that expressions such as "It's all misty" or "I can't get a grip on it", can often be precise literal unconscious descriptions from within those sensory systems, communicating unconsciously where the mind perceives a problem in handling some mental event.
Within NLP, the various senses in their role as information processors, are known as representation systems, or sensory modalities. The model itself is known as the VAKOG model (from the initial letters of the sensory-specific modalities: visual, auditory, kinesthetic, olfactory, gustatory). Since taste and smell are so closely connected, sometimes as a 4-tuple, meaning its 4 way sensory-based description. A submodality is a structural element of a sensory impression, such as its perceived location, distance, size, or other quality.
Representational systems and submodalities are seen in NLP as offering a valuable therapeutic insight (or metaphor) and potential working methods, into how the human mind internally organizes and subjectively attaches meaning to events.
According to NLP, for many practical purposes mental processing of events and memories can be treated as if performed by the five senses. For example, Einstein credited his discovery of special relativity to a mental visualization strategy of "sitting on the end of a ray of light", and many people as part of decision-making talk to themselves in their heads.
The manner in which this is done, and the effectiveness of the mental strategy employed, is stated by NLP to play a critical part in the way mental processing takes place. This observation led to the concept of a preferred representational system, the classification of people into fixed visual, auditory or kinesthetic stereotypes. This idea was later discredited and dropped within NLP by the early 1980s, in favor of the understanding that most people use all of their senses (whether consciously or unconsciously), and that whilst one system may seem to dominate, this is often contextualized - globally there is a balance that dynamically varies according to circumstance and mood.
NLP asserts that for most circumstances and most people, three of the five sensory based modes seem to dominate in mental processing:
The other two senses, gustatory (taste) and olfactory (smell), which are closely associated, often seem to be less significant in general mental processing, and are often considered jointly as one.
For this reason, one often sees the term VAK in NLP reference texts, to signify these three primary representational systems, as well as the term 4-tuple (or VAKOG) if the author wishes to include all senses including taste/smell. The same term is also known as First Access (John Grinder), or primary experience (Freud).
Putting these together, this is a very simplified example of some steps which might actually be involved in replying to a simple question such as "Do you like that dress?". The table below is useful for teaching how to identify and access each representational system in context:
| Step | Activity | Notation | What it's being used for |
|---|---|---|---|
| 1 | auditory external | Ae | Hear the question |
| 2 | visual internal | Vi | picture to oneself the meaning of the question |
| 3 | visual external | Ve | look at the dress |
| 4 | visual internal constructed | Vic | create a mental image of the dress worn by the person |
| 5 | kinesthetic internal | Ki | get an internal feeling from looking at it |
| 6 | auditory internal dialog | Aid | ask oneself 'Do I like that impression?' |
| 7 | auditory external | Ae | reply |
Logically, these or similar steps must take place somewhere in consciousness in order to cognitively make sense of the question and answer it. A sequence of this kind is known in NLP as a strategy - in this case, a functional outline of the strategy used by the mind in answering that question. In a similar way, the process leading to a panic attack of the form "I see the clock, ask myself where the kids are, imagine everything that could be happening and feel scared" might be notated as having a subjective structure: Ve → Aid → Vic → Ki, signifying that an external sight leads to internal dialog (a question), followed by internal and constructed images, leading to a feeling.
It's worth noting that usually, some of these steps (often the most important ones) occur extremely fast, and out of conscious awareness. For example, few people would ordinarily be aware that between question and even considering an answer, there must be steps in which the mind interprets and contextualizes the question itself, and steps which explore various possible strategies to be used to obtain an answer and select one to be followed. The mental occurrence of these steps is often identified by deduction following skilled observation, or by careful inquiry, although their presence is usually self-apparent to the person concerned once noticed.
A common (but not universal) style of processing in the West is shown in the attached chart, where eye flickers in specific directions often seem to tie into specific kinds of internal (mental) processing. NLP also suggests that that sometimes (again not universally), such processing is associated with sensory word use, so for example a person asked what they liked about the beach, may flick their eyes briefly in some characteristic direction (visual memory access, often upwards), and then also use words that describe it in a visual sense ("The sea looked lovely", and so on). Likewise asked about a problem, someone may look in a different direction for a while (kinesthetic access, typically downwards) and then look puzzled and say "I just can't seem to get a grip on things". Taken together, NLP suggests such eye accessing cues (1) are idiosyncratic and habitual for each person, and (2) may form significant clues as to how a person is processing or representing a problem to themselves unconsciously.
Common (but not universal) Western layout of eye accessing cues:
Eye movements to the left or right for many people seem to indicate if a memory was recalled or constructed. Thus remembering an actual image (Vr) is associated more with up-left, whilst imagining one's dream home (Vc) tends (again not universally) to be more associated with up-right.
The use of the various modalities can be identified based by learning to respond to subtle shifts in breathing, body posture, accessing cues, gestures, eye movements and language patterns such as sensory predicates.
Representational systems are also relevant since some tasks are more optimally performed within one representational system than by another. For example, within education, spelling is better learned by children who have unconsciously used a strategy of visualization, than an unconscious strategy of phonetically "sounding out". When taught to visualize, previously poor spellers can indeed be taught to improve. NLP proponents also found that pacing and leading the various cues tended to build rapport, and allowed people to communicate more effectively. Certain studies suggest that using similar representational systems to another person can help build rapport, whilst other studies have found that merely mimicking or doing so in isolation is perceived negatively.
Skinner and Stephens (2003) explored the use the model of representational systems in television marketing and communications.
Some exercises in NLP training involve learning how to observe and respond to the various cues in real time.
Originally NLP taught that people preferred one representational system over another. People could be stuck by thinking about a problem in their "preferred representational system" (PRS). Some took this idea further and categorised people as a auditory, kinesthetic, and visual thinkers (see also: learning styles) It was claimed that swifter and more effective results could be achieved by matching this preferred system. Although there is some research that supports the notion that eye movements can indicate visual and auditory (but not kinesthetic) components of thought in that moment,the existence of a preferred representational system ascertainable from external cues (an important part of original NLP theory) was discounted by research in the 1980s. Some still believe the PRS model to be important for enhancing rapport and influence. Others have de-emphasized its relevance and instead emphasize that people constantly use all representational systems. In particular, new code emphasizes individual calibration and sensory acuity, precluding such a rigidly specified model as the one described above. Responding directly to sensory experience requires an immediacy which respects the importance of context. John Grinder has stated that a representational system diagnosis lasts about 30 seconds.
In a review of research findings, Sharpley (1987) found little support for individuals to have a "preferred" representational system (PRS), whether in the choice of words or direction of eye movements, and the concept of a preferred representation system (PRS). Similarly, The National Research Committee found little support for the influence of PRS as presented in early descriptions of NLP, Frogs into Princes (1979) and Structure of Magic (1975). However, "at a meeting with Richard Bandler in Santa Cruz, California, on July 9, 1986, the [National Research Committee] influence subcommittee... was informed that PRS was no longer considered an important component. He said that NLP had been revised." (p.140) The NLP developers, Robert Dilts et al. (1980) proposed that eye movements (and sometimes gestures) correspond to accessing cues for representations systems, and connected it to specific sides in the brain.
The preferred representational systems have been used in education since the 1970s. According to a survey in 1977, 99 percent of teachers believed that a student's modality strengths and weaknesses should be considered and that a student learned more when instruction was modified to match preferred modality patterns. Pupils are diagnosed with a questionnaire as visual, auditory or kinesthetic (VAK) learners and taught according to their preferred representational system or modality. The most famous learning styles model that includes VAK is the Dunn & Dunn model, first published in 1978. The Dunn & Dunn model also includes a fourth, tactile modality, but the research that the model is based on only includes one bodily modality, usually described as kinesthetic. Michael Grinder has written about learning styles from NLPs point of view in his book Righting the Educational Conveyor Belt, giving instructions of how to make the diagnosis with NLP tools such as observing eye movements.
There was wide scientific research on modality-based instruction until the end of the 1980s. Despite the popularity of the method, assessment instruments were shown to possess poor reliabilities and the people performing the experimental task were generally not properly trained NLP practitioners. A meta-analysis synthesised data from 39 studies that represented about 3,100 subjects. With the mean effect size of 0.14, modality instruction was shown to be one of the most inefficient methods that has been used in special education. The conclusion made for educators was that substance is more important than style.
Since then, research on modality instruction has largely been concentrated to the Center for the Study of Learning and Teaching Styles at St. John's University with doctoral programmes on the Dunn & Dunn model. The developer of the method runs both the research center and a company marketing it. This research was evaluated by a British group of scholars in 2004. The conclusions about the model were unequivocal:
Despite a large and evolving research programme, forceful claims made for impact are questionable because of limitations in many of the supporting studies and the lack of independent research on the model.