← Blog

The Interstitial Journal blog · Built on evidence, not hype

What a Mood Tag Is, and What It Isn't

Naming how you feel has a genuine experimental literature behind it — better than most things in this category. Almost none of that literature is about the chart the app draws afterwards. That gap is the whole subject of this piece, because the field has quietly optimised for the part with no evidence and treated the part with evidence as a data-entry step.

By comun.ai · 17 August 2026 · ~6 min read

The one finding that justifies the feature

Matthew Lieberman and colleagues had participants view emotional images while either choosing a word for the emotion or doing a matched control task. Putting the feeling into words was associated with reduced amygdala response, alongside increased activity in right ventrolateral prefrontal cortex.¹ Moderate The effect has been studied repeatedly since under the name affect labeling.

Jared Torre and Lieberman's later review is the more useful document, because it is careful about what kind of thing this is. They characterise affect labeling as an implicit form of emotion regulation — it appears to attenuate emotional response without people intending it to, and, notably, without people believing it will.² Moderate Asked in advance, participants tend to expect naming a feeling to make it worse, or to do nothing.

Keep the size in proportion. These are laboratory studies with modest effects, and the review is explicit that the mechanism is still argued over. It supports "naming a feeling does something measurable", not "naming a feeling fixes the feeling".

The word matters more than the score

Now the part that argues against most mood-tracking design. If the active ingredient is putting the state into language, then the resolution of the language is not incidental — it is the variable.

Lisa Feldman Barrett and colleagues studied what is usually called emotion differentiation, or emotional granularity: the degree to which someone distinguishes between negative states rather than experiencing them as one undifferentiated bad. People higher in differentiation reported using more emotion-regulation strategies, particularly at high intensity.³ Moderate Todd Kashdan, Barrett and Patrick McKnight's review gathers the wider case that granularity is associated with a range of better outcomes.⁴ Moderate

Read carefully, that is largely correlational. Nobody has established that deliberately increasing your vocabulary for feelings causes the benefits, and we are not going to say they have. Early But it does something to the design question. A five-point scale from bad to good is a container that cannot hold a distinction. "Frustrated with a specific thing" and "quietly dreading a specific thing" both compress to a 2, and the 2 is what survives into next week.

James Russell's circumplex model is the reason those coarse scales exist at all: much of the variance in reported affect organises along two dimensions, roughly pleasantness and activation.⁵ Moderate That makes a two-axis picture a defensible summary of a population. It does not make it a good record of a Tuesday afternoon.

Why the timing is the other half

There is a well-established methodological reason to capture a state when you are in it rather than at the end of the day. Mihaly Csikszentmihalyi and Reed Larson set out the Experience Sampling Method precisely to escape retrospective distortion — prompt people at moments during ordinary life and record the state then.⁶ Strong as a methodological principle; it is how a large part of modern affect research is conducted.

The distortion it escapes is not a small one, and we have been through it on this blog: recall of an episode is disproportionately built from its peak and its ending rather than from its duration, a pattern documented by Daniel Kahneman and colleagues.⁷ Strong We covered what that does to a day's memory in What to Write When Nothing Happened.

So an end-of-day mood rating is not a summary of your day. It is a reading of how you feel at the end of your day, wearing the costume of a summary. A tag placed at the seam — when you stop, switch, or start — is a different measurement, taken from inside the thing.

What a week of tags does and does not tell you

Here is where we should be most careful, because this is where mood-tracking products routinely overreach.

The act of recording can itself change what is recorded. Reactivity in daily-diary and experience-sampling research has been examined directly; the general finding is that it is usually small, though it is a real consideration in study design rather than a non-issue.⁸ Early Applied to a personal log, that cuts both ways: your record is a slightly participatory instrument, and a pattern in it is not automatically a pattern in you.

Two more limits worth stating plainly:

• Sampling is not random, and yours is worse than a study's. Experience sampling prompts you; a journal waits for you to arrive. You will tag more when you are unsettled and less when absorbed, which is a bias no amount of averaging removes.
• A trend line is not a diagnosis. Clinical assessment uses validated instruments, administered and interpreted by people qualified to do it. A tag you invented on a Tuesday is not on that list, and a downward slope in an app is a prompt to talk to someone, never a conclusion.

Where the evidence stands

What this means for the tag in this app

Interstitial Journal lets you optionally name a mood alongside a timestamped line. Given the above, three design positions follow, and it is worth being explicit about them.

The mood is optional, and stays optional. A required field turns a reflective act into a compliance one and produces exactly the low-resolution filler the granularity research suggests is worth least.

The line beside the tag is the part with the evidence. Affect labeling is about language. A tag on its own is a bucket; a tag plus four words is a label. If you only do one, do the words.

The tag is a bookmark, not a score. Its most defensible use is retrieval — finding the moments later — rather than aggregation. That is the same argument we made about keeping notes in What to Keep and What to Delete: the value of a record is not obliged to arrive as a summary statistic.

None of this makes mood tagging a large intervention. It makes it a small, cheap one with a real mechanism behind the naming and no mechanism at all behind the chart — which is a fair description of most of what belongs in an interstitial journal, and why the practice stays deliberately modest.

Where the evidence ends

To be precise about the boundary. Nothing above is a treatment for anxiety, depression, or any other condition, and tagging moods is not a mental-health intervention. Affect labeling was studied as a laboratory phenomenon, not as a therapy. Emotion differentiation research describes associations in populations, not a promise about your week.

If your moods are distressing you, or the pattern you are seeing worries you, the useful next step is a qualified professional, not a better tagging habit. This is reflection, not treatment. More on our science page, and on how your notes stay private on our privacy page.