The original eye-accessing-cue chart
The model, as first taught by Bandler and Grinder in the 1970s, maps six eye positions (from the observer's point of view, looking at the person being observed):
- Up and to the left — visual remembered (recalling an image you've seen).
- Up and to the right — visual constructed (building an image you haven't seen).
- Level and to the left — auditory remembered (recalling a sound or voice).
- Level and to the right — auditory constructed (constructing a new sound).
- Down and to the left — auditory digital / internal dialogue (talking to yourself).
- Down and to the right — kinesthetic (body sensations, emotions, tactile memory).
The original framing reverses for left-handed people, which already complicates the picture: the cue depends on individual neural organisation. That qualification rarely survives translation into the strict eye-movement-as-lie-detector version that turns up in popular books.
What the research actually shows
Multiple controlled studies — most recently Wiseman et al. (2012) — have failed to find a reliable mapping between eye direction and lying, recall vs. construction, or specific sensory modalities. The strict version of the model is not empirically supported.
What is supported: eye movements during cognitive tasks correlate with processing load and with whether information is being recalled or generated. The weaker claim — that you can see a person thinking in their eye movements — is real. The stronger claim — that you can decode what they are thinking from the direction — is not.
How working NLP coaches use the cues today
Most working practitioners treat eye-accessing cues as a soft signal in calibration. A typical workflow:
- Calibrate. Ask the client questions you know the answer to ("describe your front door"). Notice where their eyes go. This builds their individual pattern, not the textbook one.
- Notice changes. Mid-conversation, if their eye-movement pattern shifts noticeably, something has changed internally — whether you can label it or not.
- Use cues as question prompts, not conclusions. "I noticed you looked up — were you picturing it?" gets useful information. "You looked up so you must be lying" is junk.
What to use instead, or alongside
Eye movements are one of many calibration signals. More reliable ones include:
- Sub-modality language. Visual people use "I see what you mean"; auditory people use "that sounds right"; kinesthetic people use "I feel that". Track the words.
- Breathing. High and shallow vs. deep and low maps roughly to visual vs. kinesthetic processing.
- Voice tempo and pitch. Faster and higher tracks visual; slower and lower tracks kinesthetic.
- Skin colour and micro-expression. Faster to read than eyes once trained.
Frequently asked questions
Do eye-accessing cues actually work?
The strict version of the model — that eye direction reliably maps to a specific sensory representation — has not replicated under controlled testing. Working practitioners use eye movements as one weak signal among many, never as a diagnostic. Treat them as something to notice and verify, not as a lookup table.
Why do my eyes move when I think?
Eye movements during recall and construction are real. They are linked to how the brain accesses memory and constructs mental imagery. The reliable claim is 'eye movements correlate with internal processing'; the contested claim is 'specific directions map to specific sensory modalities'.
Can I read someone's mind from their eye movements?
No. Even in the original NLP framing, eye-accessing cues were a calibration tool — you established a person's individual pattern before drawing any conclusion. Reading 'strangers' from eye direction is unreliable and was never the intended use.
How do NLP coaches actually use eye-accessing cues today?
Mostly as a soft signal. If a client looks up while answering a question, the coach might gently probe for visual imagery. If they look down, the coach might probe for body sensation. The eye movement is a question prompt, not a conclusion.