Descriptive Experience Sampling

What is actually happening in your mind right now? Not what you think is happening, not what you'd report in a questionnaire, not what your theory of mind tells you should be happening — what is actually there, at this precise instant, in your experience?

Russ Hurlburt has spent fifty years trying to answer this question, one random beep at a time. His method — Descriptive Experience Sampling — is disarmingly simple: wear a beeper, and when it goes off, freeze whatever was in your experience at that exact moment. Then describe it in painstaking detail during an interview. The results are, without exaggeration, some of the strangest data in all of psychology.

The Method

The beeper goes off at random intervals throughout the day. At each beep, you jot a note about what was in your experience at the "last undisturbed moment" before the beep interrupted you. Then, typically the next day, you sit down with a trained DES interviewer and describe each moment in exhaustive detail.¹

The critical word is "trained." Hurlburt's key insight — hard-won over decades — is that people are terrible at describing their inner experience on the first try. Not because they're lying, but because they don't have the vocabulary and they don't know what counts. Lena, a university student who contacted Hurlburt in 2020 convinced she had constant internal monologue, produced a classic first-day report: a long, wandering narrative mixing context, speculation, memory, emotional reaction, and actual experience into an undifferentiated stream. The commentary on her transcript is blunt: this is "a very typical first-day-first-sample DES description." Most people can't distinguish experience from context, or limit themselves to the moment of the beep, without practice.²

The training takes days. Over multiple sampling sessions, participants learn two fundamental skills: distinguishing experience from everything else (context, background, interpretation), and limiting themselves to the moment of the beep and no other time. By day three or four, the descriptions sharpen dramatically. This is why questionnaire-based measures of inner experience produce wildly different results — the respondent hasn't been trained to notice what they're actually reporting on.

The Five Frequent Phenomena

Across hundreds of participants and thousands of beeps, Hurlburt's group identified five phenomena that account for most of inner experience. The canonical statistics come from Heavey & Hurlburt (2008): 30 university students, 10 randomly sampled beeps each.³

Inner speaking (~26% of samples, range 0–75%). You are saying something to yourself in your own voice. Not metaphorically — you experience the words, the prosody, the pacing. But here's the shock: most people believe inner speech is constant, running-monologue style. On questionnaires, they report it at 68%. DES catches them at 26%. The gap is not subtle — it's a factor of nearly three. When Hurlburt asks people to recall exactly what they were saying at the moment of the beep, many discover they weren't actually saying anything at all.³⁴

Inner seeing (~34%, range 0–90%). Visual imagery — seeing something in your mind's eye that isn't physically present. Some people experience this almost constantly; others never do (see Mental Imagery on the aphantasia spectrum). Lena, who believed she was primarily a verbal thinker, turned out to have inner seeing in 54% of her samples and inner speaking in only 19%.²

Unsymbolized thinking (~22%, range 0–80%). This is the most controversial phenomenon — and the one most people can't recognize without training. You are thinking something specific, with definite content, but there are no words, no images, no symbols of any kind. You know what you're thinking — you could articulate it if asked — but at the moment of the beep, the thought has no experiential vehicle. Many philosophers consider this impossible. DES participants report it routinely once they learn to notice it.³

Feeling (~26%, range 0–90%). An emotional experience with an experiential quality — not just the cognitive label "I am angry" but the felt texture of anger. DES distinguishes this from "emotional process" (your body doing anger-things that you're not attending to). At the moment of the beep, the feeling is either in your experience or it isn't, regardless of what your body is doing.

Sensory awareness (~22%, range 0–100%). Paying attention to a sensory experience for its own sake — the warmth of sunlight, the texture of fabric — rather than using it instrumentally. One participant reported sensory awareness in 100% of samples. Nine reported it in 0%.

The ranges are the most important numbers in the table. Inner speaking varies from 0% to 75% across individuals. Some people never experience inner speech. Others experience it three-quarters of the time. These are not pathological cases — they are normal university students leading normal lives, completely unaware that their inner experience differs radically from the person sitting next to them in class. This connects directly to the introspection problem: we assume everyone's mind works like ours, and we're wrong.

The Inner Experience of... Series

Starting in 2020, Hurlburt made DES radically transparent. The "Inner Experience of..." series on his website documents 14 participants with full YouTube interview videos and annotated PDF transcripts — every question, every hesitation, every correction, published for anyone to scrutinize.²

The participants span a remarkable range. Ryan Langdon, whose blog post "Today I Learned That Not Everyone Has An Internal Monologue" went viral in 2020 and triggered the whole public series. Sadie Dingfelder, a journalist with prosopagnosia, stereoblindness, aphantasia, and severely deficient autobiographical memory. Michael Pollan, the bestselling author, whom Hurlburt concluded had "very little inner experience" at beeped moments — Pollan disagreed, and the disagreement itself is published. Mel May, an Australian video producer whose inner experience was "close to being nothing." Kerry U., who had constant inner speech (the contrast case to Mel).

The Lena case is the most systematically documented. Krumm & Hurlburt published it as a "complete, unabridged, pre-registered" investigation — the first in DES history where every step was made public in near-real-time.² Eleven sampling days produced 37 analysed beeps. Lena, who had contacted Hurlburt because of the internal monologue debate, discovered three things: she didn't have constant inner speech (zero examples of monologue-style narration across all 37 samples), she had frequent visual imagery she hadn't known about, and she often experienced two simultaneous "centers of gravity" — distinct experiential foci that were neither the same experience nor two separate ones, but something in between that existing vocabulary struggles to capture.

What DES Reveals That Nothing Else Does

Only 3% of reading involves inner narration. Brouwers et al. (2018) beeped 16 college students while they read Fitzgerald and Hemingway. Almost nobody was "hearing" the words in their head. This contradicts one of the most universal self-reports about reading — the sense that you hear an inner voice narrating the text. DES suggests this is a retrospective confabulation: when asked "do you hear an inner voice while reading?", people say yes because it seems like they should, not because they've caught themselves doing it.⁵

Autistic inner experience is radically visual. Hurlburt, Happé, and Frith (1994) studied three adults with Asperger syndrome. All three reported thinking almost exclusively in images — up to 100% visual imagery in some participants — with essentially no inner speech. This confirmed Temple Grandin's famous self-report but through random sampling rather than retrospective self-assessment.⁶

Bulimia correlates with fragmented multiplicity. Jones-Forrester (2009) studied 24 women with bulimia. All 24 — every single one — showed "fragmented multiplicity": multiple simultaneous experience fragments, often a dozen or more, none fully formed, often with "tails" (elements lingering after leaving direct awareness). One participant, while watching TV, had two simultaneous inner voice streams — a louder one in the front of her head and a quieter one in the back — that, assembled, would form a coherent sentence, but at the moment of the beep she didn't experience the sentence, only the fragments.

Borderline personality disrupts figure/ground. Fran, a bank teller studied in Hurlburt's 1990 book, showed no figure/ground distinction in perception — she could see both the faces AND the vase simultaneously in ambiguous-figure illusions where most people can only toggle between them.

Some people may have no inner experience at all. One of Hurlburt's 1990 participants, a man with schizophrenia, appeared to have no detectable inner experience at beeped moments. Not suppressed or inaccessible — genuinely absent. This is the extreme case of what many normal participants discover on their first sampling day: the beeper goes off and there is simply nothing there.

The Questionnaire Gap

The most damning statistical result in DES research is the comparison between questionnaire-based and DES-based frequency estimates. Hurlburt et al. (2022) compared the Nevada Inner Experience Questionnaire (NIEQ), Brinthaupt's Self-Talk Scale, and DES directly. The results were, in their word, "irreconcilably discrepant."⁴

People overestimate the frequency of virtually everything in their inner life — but especially inner speaking (by 42 percentage points) and feelings (by 48 points). Hurlburt's interpretation: without DES training, people don't have adequate understanding of their own inner experience to answer questions about it. They report what they believe should be there, not what's actually there. This is introspection's unreliability quantified.

The Criticism

Eric Schwitzgebel, who spent an entire book (2007) doing DES alongside Hurlburt with a participant named Melanie, remains the primary skeptic. His position: even with DES's safeguards, first-person reporting is "too subject to distortion." The imprecision of introspective concepts and the shifting, ephemeral nature of the phenomena render DES reports open to multiple interpretations. He points to the dreaming-in-color problem: in the 1950s people reported dreaming in black-and-white; after color TV, they reported color dreams. We probably just don't know.

The structural critiques are real: DES is essentially a one-lab operation. The method requires years of interviewer training, making independent replication nearly impossible. Most case studies are N=1. The largest nomothetic study was N=30. When Gunter (2011) tried to replicate Hurlburt's finding that depression increases unsymbolized thinking, she failed.

But the strongest results — the questionnaire gap, the 3% reading figure, the autism finding — are so large they survive any reasonable amount of methodological noise. A factor-of-three overestimation of inner speech isn't a subtle effect that could be explained by demand characteristics. Either people are catastrophically wrong about their own inner lives (Hurlburt's position), or DES systematically distorts what it measures (Schwitzgebel's position). Both possibilities are disturbing.

What This Means

DES's deepest implication isn't about any specific finding — it's about the epistemic status of first-person reports. Every theory of consciousness, every model of cognition, every philosophical argument about qualia depends on someone being able to say what their experience is like. DES suggests this ability is far worse than anyone assumed, but also that it can be trained — that with practice, people can learn to notice things about their own minds that decades of living never revealed.

The fact that it took until 2021 for the first complete, unabridged DES case study to be published — after fifty years of the method existing — says something about how protective the field has been of its raw data. Hurlburt's recent pivot toward radical transparency (the IEO video series, the pre-registered Lena study) is a bet that showing the actual process of investigating inner experience, warts and all, will be more convincing than any statistical summary. Whether he's right depends on whether anyone else can learn to do what he does — and so far, that remains an open question.