«BY GEORGIOS DIAMANTOPOULOS A THESIS SUBMITTED TO THE UNIVERSITY OF BIRMINGHAM FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF ELECTRONIC, ...»
“What we noticed is that different people actually think differently, and that these differences correspond to the three principal senses: vision, hearing, and feeling – which we call kinaesthetics2. When you make initial contact with a person s/he will probably be thinking in one of these three main representational systems. Internally s/he will either be generating visual images, having feelings, or talking to themselves and hearing sounds.” According to Dilts and DeLozier (2000, p. 1097), “the term representational systems refers to the neurological mechanisms behind the five senses” and thus, five representational systems may be defined, each one corresponding to one of our senses: visual, auditory, kinaesthetic, olfactory and gustatory (VAKOG). The representational system is different to the “lead system”, which is the sensory system that the person uses to initiate the search for the representation of the According to Dilts and DeLozier (2000), “’kinaesthetic’ is a term used in NLP to refer to feelings and body 2 sensations […] it is used to encompass all types of feelings including tactile, proprioceptive and visceral”.
15 experience; for example, a search for a visual representation may be initiated through an auditory or kinaesthetic representation. There is also the distinction of the primary representational system (PRS), which is introduced and loosely defined in the original NLP texts (Bandler and Grinder, 1975; Grinder and Bandler, 1976; Bandler and Grinder, 1979); an appropriate definition is found in Dilts and DeLozier (2000, p. 1102): “in NLP, a person is said to have a primary representational system when that person values or uses one of his or her senses over the others in order to process and organise his or her experience of the world.” Using this notion of representational systems, they suggested that observable body language cues such as eye-movements, voice tone and tempo, body posture, gestures and breathing patterns indicate which representational system the person is currently employing (Bandler and Grinder, 1979).
THE NLP EYE-ACCESSING CUES MODEL
The NLP Eye-Accessing Cues (EAC) model was first introduced by Grinder, DeLozier and Bandler (1977) and further refined by Bandler and Grinder (1979). The EAC model suggests that nonvisual eye-movements (i.e. eye-movements that are not concerned with the visual pursuit of an object in the environment) indicate which representational system a person is currently using.
According to the model, such eye-movement patterns are observed in all individuals regardless of handedness in an idiosyncratic fashion. In other words, each individual ought to display an eye-movement pattern correlated to the modality they are currently accessing that is consistent to them. An explicit generalisation is offered for normally-organised right-handed people (Bandler and Grinder, 1979, p. 25; Figure 1 below). It is a generalisation because the patterns are said to be idiosyncratic and in this sense, the pattern is generalised to normally-organised righthanded people.
In simple terms, assuming a normally-organised right-handed subject that conforms to the generalised pattern offered in Figure 1 below, the EAC predicts that when, for example, the subject is accessing a visual memory, s/he is going to look up and right (their left). Thus, if the subject was asked a question such as “do you know when tomatoes are ripe?” and an eyemovement up and to the right is observed, that s/he accessed a visual memory in answering the 16 question, perhaps that of a ripe tomato. In practice, responses are usually much more complex but this crude example serves to illustrate the concept of the EAC model to the unfamiliar reader.
FIGURE 1: A GENERALIZATION OF THE EYE-ACCESSING CUES MODEL FOR A NORMALLY-ORGANISED
RIGHT-HANDED PERSON (BANDLER AND GRINDER, 1979). IT IS A GENERALISATION BECAUSE THE
PATTERNS ARE SAID TO BE IDIOSYNCRATIC AND IN THIS SENSE, THE PATTERN IS GENERALISED TO
THIS PARTICULAR GROUP OF PEOPLE.
It can be said that the EAC model is a core component of NLP since it forms the basis of several advanced NLP techniques such as strategy elicitation and installation (Bandler and Grinder,
1979) which are said to have very practical uses. For example, Malloy (1987) found that pupils significantly improved their spelling ability if they looked up and to the right (from the observer’s point of view) at the same time as visualising the word, versus visualising the word but looking in another direction and not visualising the word.
A very important part of the EAC model as presented by Bandler and Grinder (1979), is the process which elicits the eye-movements in question. This process, termed as transderivational search occurs when the subject recovers the deep structure from the surface structure, as shown in Figure 2.
The terms deep structure and surface structure were coined by Chomsky (1965) though a clear definition is elusive in both Chomsky’s and the original NLP texts. A fairly comprehensive explanation can be found in Dilts and DeLozier (2000). Briefly explained, the deep structure consists of thoughts and ideas and their linguistic expression is the surface structure; derivation is a series of transformations which connects the deep structure with the surface structure (Bandler and Grinder 1975, p. 29) and hence the transderivational search. Chomsky (1965) originally used these terms to describe linguistic processes but Bandler and Grinder implicitly 17 extended these notions to neurological processes related to our sensory experience (Dilts and DeLozier 2000). Bandler and Grinder (1975) describe three transformative processes (deletion, distortion and generalisation) that are reflected both in the linguistic and the mental representation of the person’s experience. Finally, “transderivational search is the process of accessing the meaning, which is equivalent to some set of images, feelings or sounds that are associated to that word” (Bandler and Grinder, 1979, p. 15).
Thus, transderivational search is the process that elicits the eye-movements that the EAC model focuses on. Thus, the first and foremost challenge in directly examining the EAC model as the following studies have done, is identifying an experimental methodology for eliciting eyemovements that follows a specific and precise definition of transderivational search. As useful as the above definitions may be, they require further refinement before they can be used experimentally. The methodology also needs to consistently achieve predictable responses in all instances and for all subjects. For the specific purpose of investigating the EAC model, it ought to recover visual, auditory, kinaesthetic and optionally olfactory and gustatory representations.
FIGURE 2: ELICITATION OF EYE-MOVEMENTS IN THE EAC MODEL AND TRANSDERIVATIONAL SEARCH.
PAST EYE-ACCESSING CUES MODEL RESEARCHOf the relatively few academic studies of NLP, a large percentage of them have been concerned with the EAC model and the closely related notion of the primary representational system (PRS) introduced earlier. In this section, past research literature is reviewed and critiqued, with a sole focus on the EAC model. References to studies of the PRS will only be made where
Examination of the issues that revolve around handedness is beyond the scope of this review; the generic form of the model is assumed here, which, as stated earlier, suggests that patterns emerge in all individuals regardless of handedness.
BRIEF OVERVIEW OF OBJECTIVES AND REPORTED RESULTS OF EAC STUDIESIn the context of this review it will be said that a study has shown partial support for the EAC model if any of its results are statistically significant and consistent with the EAC model. Further, a study will be said to be unsupportive or to have shown no support for the model if none of its results are statistically significant or if none of its statistically significant results are consistent with the EAC model.
Thomason et al. (1980) attempted to test the EAC model hypothesis using questions to elicit visual, auditory and kinaesthetic (VAK) representations; their study was unsupportive of the model and was criticised by Beck and Beck (1984). Elich et al. (1985) used questions to elicit VAK representations and attempted to correlate eye movements and verbal predicates with question modality (interview-style); their conclusion was unsupportive of the model. Buckner and Reese (1987) asked subjects to report on VAK components of pleasant thoughts and found partial support for the model. Another test by Baddeley and Predebon (1991) correlated eyemovements with the corresponding verbal-report of the subjects’ subjective experience and also found partial support. Burke et al. (2003) tested the relation between eye-movements and visual-kinaesthetic-gustatory (VKG) tasks both as hypothesised by the NLP model and idiosyncratically and found support for the idiosyncratic hypothesis.
Farmer et al. (1985) used recall of real stimuli and found no support for the model; a similar study was repeated by Wertheim et al. (1986) and found partial support. Dooley and Farmer (1988) repeated the experiment of Farmer et al. (1985) with aphasic subjects and found partial support.
Cheney et al. (1982) tested the relationship between eye-movements and reported imagery with the use of a questionnaire on vividness (Sheehan 1967); the results were unsupportive of the
In an interview-style study (Ellickson 1983), the interviewers attempted to match the subjects’ representational system, as determined in real-time by their eye-movements, through verbal predicates and tested the effect of predicate matching on perceived counsellor empathy.
Falzett (1981) used eye-movements to assess the PRS and determine the outcome of matching it through verbal predicates in counselling. Gumm et al. (1982) attempted to examine the agreement in the determination of the PRS using eye-movements, verbal predicates and selfreport. Sandhu (1991) tested whether the PRS can be reliably determined from eye-movements by comparing its assessment from eye-movements, verbal predicates and self-report; he found no support for the EAC model.
Table 1 below is a summary of all the past research studies relevant to the EAC model.
EYE-MOVEMENT ELICITATIONUnknown questions were used by Thomason et al. (1980) and Poffel and Cross (1985); Falzett (1981) used questions from an unpublished doctoral thesis that are not reproduced in his publication. Gumm et al. (1982) used twenty questions to provide the subject with a variety of “cognitive tasks”. It is unknown what exactly is meant by “cognitive tasks” in this case; it is likely that the tasks were unrelated to direct elicitation of sensory representations as other research of the time was concerned with generic mental tasks (e.g. Ehrlichman et al., 1974). Ellickson (1983) employed six “stimulus cues” during interviews that were apparently designed to elicit eye movements and neither the design criteria nor the stimulus cues are included.
Buckner and Reese (1987) asked their subjects to “think in silence of a single pleasant thought or memory” and after ten seconds, the interviewer asked the subject to report any VAK components; once again, the validity of this methodology in eliciting eye-movements as described above was not examined by Buckner and Reese (1987).
Eye-movement elicitation methodology was only sparsely informed by other research; the only instances are the papers by Cheney et al. (1982) and Elich et al. (1985) that utilised the questionnaire on mental imagery developed by Sheehan (1967). However, its relevance and 20 validity in the investigation of the EAC model was not discussed nor tested; even if the questionnaire is successful in eliciting eye-movements, it is not appropriate for this purpose as it explores the different properties of objects within a single representation (e.g. the colour of the dishes and the food on the breakfast table).
Sandhu (1991) was the first to note and give weight to the importance of “stressed recalls” and proposes that the subject’s eyes will shift in potentially meaningful ways only when they “think hard” to answer the question. Despite this observation, the questions taken from the Sandhu PRS inventory are neither reproduced nor published in a journal publication elsewhere. Further, the inventory sample provided is arithmetic and not relevant to any one sensory modality.
Regardless of the particular method, the examples above highlight the ad-hoc selection of the eye-movement elicitation methodology and the implicit assumption that the respective methodology is equivalent of the transderivational search and consequently examines the desired eye-movements; no formal pilot studies were conducted and minimal emphasis was given on this pivotal aspect.
Baddeley and Predebon (1991) provided the full inventory of questions but there was another fundamental flaw. Duke (1968) found that a “complex” question will elicit a series of eyemovements. However, the more complex the question, the more difficult it is to isolate the cognitive process at work. Thus, questions that are too simple may not elicit any eye-movements and questions that are too complex may elicit too many eye-movements and/or cognitive processes for any useful distinctions to be made. More specifically, in the case of such complex questions, there is no guarantee that the representational system accessed by the subject is the same as intended by the author of the question. For example in a question that appears in Baddeley and Predebon (1991), “What colour are the walls in your bathroom?”, which is reported as “visually remembered”, the subject could retrieve a memory of their bathroom, kinaesthetically, e.g. by remembering the feeling of sinking into warm bath water (also reported by Beck and Beck, 1984). As mentioned earlier, in the EAC model, this is termed as the “lead system” (Bandler and Grinder, 1979, p. 28).