Benjamin Sturgeon

23 April 2026

Whatever I Was Doing in 2020, I Should Do Again (Continued)

Day 23 of Inkhaven: 30 Days of Posts

In my previous post, I focused on the content of my posts in 2020 and what led to all of the happy feelings that I had then. This leaves the other big question from the data which is what were the factors behind the huge drops in scores in 2024 and 2025.

One interesting thing to observe is how my language changed during the 2025 and 2026 period. In particular, the kinds of words changed completely. In 2020 I used a great deal of words related to decisions, and committing to things. By 2026 usage of these phrases collapsed. Here is the precise data along with additional examples:

Self-talk grammar (per 1000 words, annual averages):

CategoryExamples20202026Change
Commitment"I'll, I will, I'm going to"6.211.71Dropped to less than a third
Uncertainty"idk, I guess, I'm not sure, maybe"2.769.51More than tripled
Warmth"I love, grateful, wonderful, beautiful"0.500.00Went from infrequent usage to never being used

Specific phrases (per 10,000 words):

Phrase20202026Change
"honestly"5.534.3More than 6×
"I feel like"3.816.9Almost 5×
"idk"015.3Went from 0 to relatively common

The most important part of this I think is the part around agency and being able to make decisions. At the time in 2020 I remember explicitly cultivating a practice of making decisions for myself. In 2026, the process of making a decision and committing to something has become almost painful to me. It feels like there is a lot more fear about how things could go badly, rather than a sense of freedom and play in the options that are available.

The question is what are the causes of this. I'd say the most important piece is burnout and stress management. The low scores mostly come from entries where I went into my journal and simply copied and pasted words over and over again until I hit my 750 words. Taking the ritual I had developed for myself and simply walking through the performance so that I could get through it.

There is something fundamentally eerie and sad about doing this.


This chart shows the number of words per paragraph over time, and a clear collapse in the length of paragraph. This is reflected in the writing by their content shifting to fragmented sentences, jumping from idea to idea rather than building theses, and simply logging the inner emotional landscape rather than building towards something coherent.

This can be seen in the fact that through 2020-2022, the most recurrent word in a given entry would appear in over 60% of the paragraphs, whereas in 2025-2026 it would appear in only 34% of paragraphs. The earlier posts tended to be about a particular thing, while the later posts tend to jump across many different ideas.

Paragraph length over time

This reflects a shift in the nature of what was being recorded, although it is also naturally downstream of there being more paragraphs that fewer of them would contain a particular word.

The number of key topics also increased by roughly 2.5x going from 28 topic anchors to 70 in 2025-2026. This means that I was thinking about a much wider array of things. This helps us to build a clearer thesis. The 2025-2026 period wasn't the outlier, the intense focus of the 2020 period was what made it possible to feel as content and happy as I did.

Of course, the conditions of covid made this easier, not having any obligations, having a tiny social circle, and mostly doing the same stuff each day. But what is remarkable is how that reflected in my own happiness. This might not be a universal phenomenon, but in my case the isolation and focus led to a huge increase in my happiness.