A client sits across from you in the first session. You ask what brings them in. They say: "I just feel bad all the time. Like, not okay." You know this answer. You've heard it hundreds of times. And you know it tells you almost nothing — not about what to track, not about what's dysregulated, not about where to start. The presenting problem isn't that they're suffering. It's that they can't yet tell you what kind. That gap is the clinical triage problem this post is about.
The Problem With "I Feel Bad"
Lisa Feldman Barrett's constructed-emotion theory reframes how we understand what a client means when they say "I feel bad." In Barrett's model, emotions are not universal readouts of discrete states — they are predictions: the brain's best guess at what's happening internally, built from past experience and the conceptual vocabulary available to it. Emotional granularity, in this framework, is the precision of that prediction. A low-granularity person collapses a wide range of internal states into undifferentiated "bad" — not because they're unaware, but because the joint between interoceptive signal and emotional concept is imprecise. The brain can't produce a more specific prediction than the concepts it has been given.
For trauma survivors, this collapse is not a vocabulary problem — it is a structural one. Trauma disrupts the interoceptive-conceptual joint that makes specific labeling possible. Bessel van der Kolk's work on traumatic stress documents how the body's internal reporting system becomes unreliable: signals arrive, but the pathway between sensation and meaning has been distorted by threat responses, dissociation, and years of learned suppression. Resmaa Menakem frames it even more directly: the body holds what the mind could not process, and what is held cannot be named until it can first be felt.
The clinical implication is under-appreciated: "I feel bad" is not a starting point for emotional work. It is evidence that the client's interoceptive-conceptual system is not yet producing usable signal. Until it does, triage is guesswork.
Why You Can't Teach Vocabulary Before Building the Signal
The most common clinical error in this domain is sequencing. Practitioners who understand the value of emotional granularity often move directly to vocabulary building — offering feeling wheels, psychoeducation about the emotion families, lists of words. The intention is sound. The timing is often wrong.
Pat Ogden's Sensorimotor Psychotherapy framework makes the sequencing explicit: working with the body's sensorimotor experience is the prerequisite for working with emotional meaning, not the other way around. If a client cannot reliably attend to their body's signals — cannot distinguish tightness from trembling, heat from weight, activation from collapse — then adding vocabulary to the top of that system is adding labels to a blurry image. The label doesn't clarify the image. You need to sharpen the image first.
Three questions help practitioners assess whether a client's interoceptive capacity is ready to support differentiation work. Can the client locate sensation in the body right now, in session, without significant prompting? Can they distinguish between two similar sensations — tightness in the throat versus tightness in the chest — when asked to attend to each? Can they track a sensation's change over time within a single session, noticing when it shifts or softens? If the answer to all three is no, the work is interoceptive capacity-building first. Granularity comes after.
How Differentiation Actually Works in the Room
Once interoceptive capacity is present, differentiation is taught through contrast — not through concept. The practitioner doesn't explain what shame is and then ask the client to find it. They help the client notice two states that are easily confused, feel into the difference between them in the body, and use language to anchor what they've discovered. The vocabulary emerges from the experience, rather than being layered onto it.
Four differentiation pairs are particularly high-yield in trauma work, because these are the pairs most commonly collapsed by clients with dysregulation histories — and most consequential when collapsed:
Shame vs. guilt. Shame is located in the face and throat; it carries an urge to disappear or shrink. Guilt tends to land in the chest or gut; it orients toward repair. A client who calls everything "guilt" and cannot access shame is often avoiding accountability for the wrong reason. A client who calls everything "shame" and cannot access guilt may be unable to initiate repair. The difference matters clinically.
Anger vs. fear or terror. Both can feel like activation, urgency, or heat. But anger tends toward forward orientation — toward the threat. Fear moves away from it. Terror freezes. A client who can only report "I'm mad" when they are, in fact, terrified is using anger as a more tolerable cover state. Helping them feel the difference — in jaw versus legs, in heat versus cold — opens the clinical work.
Grief vs. loneliness. Grief is relational but retrospective — it carries loss of something. Loneliness is prospective and present-tense — it carries absence of something needed now. Collapsing these leads to different clinical errors: treating current relational need as a wound to process, or treating unresolved loss as a present-state deficiency to fill.
Numbness vs. calm. This pair trips up clients and practitioners alike. Both involve low activation and low distress. But numbness has a quality of distance or flat affect; calm has presence and groundedness. A client who reports feeling "calm" but who looks disconnected or glassy may be reporting the experience of numbness with the vocabulary of calm. Somatic tools are essential here: breath quality, postural aliveness, eye contact, and vocal tone can help distinguish them.
Somatic tools that support this work include titration (tracking tiny increments of sensation, rather than going toward the full experience), pendulation (moving attention between a difficult sensation and a resource state, building tolerance through contrast), and dual awareness (maintaining simultaneous contact with present-tense safety and the internal experience being explored). Common mistakes: naming before tracking — asking "what is this feeling?" before the client has attended to where and how it's showing up in the body; and pushing differentiation when the client is flooded, which collapses the window of tolerance further rather than expanding it.
What Granularity Unlocks — IFS, Parts Work, and Affect Regulation
Parts-based modalities depend on distinguishable internal states. In IFS, the core therapeutic question — "can you find that part in your body?" — only generates useful information if the client can distinguish between parts. The part that is angry and the part underneath it that is terrified are two distinct internal experiences, each with its own body location, its own orienting quality, its own implicit need. Without differentiation, the client can only report "there's something bad in here." The therapy cannot move.
Granularity also enables early-detection affect regulation — catching dysregulation at the edge of the window of tolerance rather than after a full flood. A client who can notice "I'm starting to feel that particular tightness that precedes shutdown" has a regulatory option. A client who only notices "I feel bad" discovers it when they're already out of the window. The earlier the detection, the more available the prefrontal cortex — and the more adaptive the regulatory response.
Barrett's construct of affect complexity — the capacity to hold multiple differentiated affective states simultaneously — is the downstream product of sustained granularity work. Todd Kashdan and colleagues have documented that higher affect complexity predicts better outcomes across a range of psychological health indicators, including resilience, recovery time after stressors, and reduced reactivity. The clinical implication is direct: building granularity is not a side skill. It is a core affect-regulation intervention with measurable outcomes.
A Practitioner Note
The shift that changed how I did this work was moving from asking "can you name it?" to asking "can you feel the difference?" The first question assumes the signal is already specific enough to match to a word. The second question accepts that the signal may still be blurry — and works with the blurriness directly.
When I stopped trying to bring vocabulary to clients and started helping them feel contrasts in their bodies — this versus that, here versus there, when activated versus when settled — the language started arriving on its own. Not because I taught them words, but because they had something real to attach words to. The word "shame" means something different once you've felt the particular quality of wanting to disappear. You don't need to explain it. You just need to help them find it.
This work requires patience because it requires sequencing. You can't rush interoception. You can't teach contrast before the client can attend to a single sensation. And you can't differentiate when someone is flooded — you can only wait, regulate, and return. But the patience is itself precision. Taking time to build the signal before building the vocabulary is not a slow path to the work. It is the work. And once the granularity is there, once a client can distinguish between their shame and their grief and their terror and their calm — the rest of the therapeutic process has something solid to move on. The co-regulating relationship is the mechanism. The vocabulary is what it builds.
"You can't teach a client to name what's happening if the body hasn't yet learned to send a specific enough signal. First you build the interoception. Then you build the vocabulary. Then — and only then — does the word actually correspond to something real."