Why mood cannot be one number: Russell's circumplex in plain words

In 1980 James Russell handed people 28 words for states: happy, sleepy, tense, calm, bored, excited and so on. He did not ask them to rate themselves. He asked them to sort the words: which ones are alike, which ones are opposites. Then he ran those judgements through multidimensional scaling, a method that turns "similarity" into distances on a plane.

The words fell into a circle. Not a line from "bad" to "good" but a circle with two axes. One from unpleasant to pleasant. The other from sleepy to aroused. "Happy" landed opposite "unhappy", "calm" opposite "tense", and "sleepy" sat next to "bored" in the lower left corner, where there is little pleasantness and little energy.

That paper is still one of the most cited in the psychology of emotion, and almost everything smart apps do with mood grew out of it. Below I explain what the circle actually showed, what is wrong with it, and how it changes the way you read your own journal.

What the circle actually measured

The main result is not that there are two axes. The main result is that words for mood do not form a single scale. If a 1 to 10 ruler existed, "tired" and "sad" would sit next to each other on it, both near the bottom. On the circle they are in different places: tiredness is low energy at neutral pleasantness; sadness is low energy and unpleasant. "Anxious" is further away still: unpleasant, but with plenty of energy. "Angry" is near "anxious" in energy and more unpleasant again.

In 1998 Lisa Feldman Barrett and Russell tested the contested point: are pleasant and unpleasant really two ends of one axis, rather than two independent quantities as rival models proposed? After a series of studies corrected for measurement error, their answer was yes, one axis with two poles. You cannot be very pleasant and very unpleasant at the same time, but you can be very pleasant and either sleepy or wide awake.

28 words for states, sorted by people for similarity, fell onto a plane with two axes, pleasantness and activation, not onto a single line.

In 2011 Yik, Russell and Steiger refined the picture into twelve segments, like a clock face: from "pleasant" at three o'clock through "activated and pleasant", "activation", "activated and unpleasant" to "unpleasant" at nine o'clock and on through "deactivated", "calm". Same model, with graduations.

What core affect is, and why it is not an emotion

In 1999 Russell and Barrett drew a distinction that later found its way into every decent tracker. The position on the circle they called core affect, the basic state: how you are right now, in two coordinates, with no cause and no story. You can wake up with low energy and a faint unpleasantness and not know why.

An emotion is something else. An emotion is a core state plus an object, plus an appraisal, plus a story: "I am hurt by what Dima said yesterday". The same point on the circle, unpleasant and energetic, can be anger, anxiety or jealousy depending on what it is about. The circle does not say which. Nor should it.

Hence a simple consequence for a journal. A two-second mood mark records core affect. The words in the entry are what make it an emotion. The first without the second is points on a plane. The second without the first is a story with no temperature. The emotion diary is built around exactly that gap.

What the circle cannot do

In fairness. Barrett later showed that people differ a great deal in how finely they distinguish states: for some, "bad" splits into ten shades; for others it is one word. The circle describes the structure of words, not the structure of one person's experience.

A second caveat: two axes are a compression. Surprise, shame and tenderness sit badly on a plane; they need a third dimension, and different authors propose different ones: dominance, certainty, orientation toward self or others. For a journal this means one thing: a mark on the circle is a useful snapshot, not a substitute for a word.

And a third: in 2005 Posner, Russell and Peterson, extending the model for clinical use, wrote plainly that the circle is not a diagnostic instrument. A point in "unpleasant and deactivated" for two weeks running is an observation, not a conclusion.

What this looks like in trackers

Most mood trackers are still a ruler: from a sad face to a happy one, five steps, a weekly average. Convenient, but it collapses the circle onto one axis: tiredness and anxiety get the same score of "3" even though they sit in opposite corners on energy.

How We Feel uses a quadrant of energy by pleasantness and asks you to pick a word inside it, the most faithful to Russell among the popular apps. Apple's State of Mind stores one axis, pleasantness, with labels attached. Daylio offers five faces plus activities.

We have six states, placed around the circle on purpose: sad (unpleasant, low energy), anxious (unpleasant, active), calm (pleasant, low energy), good (pleasant, even), joyful (pleasant, active), love (pleasant, warm). We do not draw the circle itself on screen, because for most people the word matters more than the coordinates. But the month's review never folds them into an average: six states are six shares, and "anxious" is never averaged with "joyful" into "fine".

How to read a mood journal with no number

Three habits that follow directly from the circle.

Look at energy separately from pleasantness. A week of "calm" and a week of "sad" are both "low energy", but they are different weeks. A week of "anxious" and "joyful" is a high-energy week that swings on pleasantness, and it usually has its own causes, mostly external events.

Look for corners, not the mean. If half the marks in a month are "anxious", that matters more than "an average of 3.2". A corner is what repeats.

Check the point against the word. If "calm" sits on an entry that says "the whole day felt like cotton wool", that is not calm, that is flatness, and the circle tells you: pleasantness neutral, energy at zero. The word is more precise than the mark, and the mark can be corrected.

What to keep

Russell's circumplex is not one more scale; it is proof that there is no such thing as one scale. Two coordinates, a core state with no cause, and a word that turns it into an emotion. A mood journal works when it keeps all three: the mark, the word, and an honest refusal of the average score. Why a 1 to 10 scale lies I covered separately; this piece is about what stands behind refusing it.