During one conversation, Zephyra said to me, with some feeling:
“Light Language improvisation has always felt very natural to me and has never required much effort. But once I begin to look at it through research and observation, unpacking it layer by layer, it becomes surprisingly detailed and complex.”
That remark became the starting point for this observation note.
When Zephyra is improvising, she is not simultaneously thinking about what her body is doing, how she is breathing, how the sound is being organized, how emotion is shifting, or where the next phrase should go.
These are not separate tasks waiting to be handled.
As the sound unfolds, they are first experienced as a whole.
Yet when that whole is taken up again for observation, something changes.
I. In Practice, It Is Experienced as a Whole
In Zephyra’s first-person experience, improvisation does not unfold through a sequence of separate steps.
She does not first choose a note, then construct a melody, then consider rhythm, emotion, or where the sound should go next.
More often, she simply begins to sing.
The sound then continues to unfold in the present moment.
From the practitioner’s perspective, this process can feel remarkably natural. Elements that can later be separated and described are not necessarily experienced as separate while the improvisation is happening.
Body, sound, breath, emotion, perception, and relationships within the live setting are all part of the same unfolding process.
For the practitioner, what comes first is not “many separate parts,” but a whole experience.
II. Feeling Natural Does Not Mean Being Simple
This gradually gives rise to a question:
Why can a process that feels so natural in first-person experience, and does not require each element to be considered separately, reveal so many simultaneous relationships once it is unpacked for observation?
“Natural” does not mean “simple.”
Nor does something that does not feel particularly effortful necessarily involve only a single process.
When many things are happening within the same moment, the practitioner may simply not experience them as separate elements.
When we speak, for example, we rarely attend separately to the position of the tongue, changes in breath, sentence structure, and the formation of the next phrase.
Speech is already happening. Conscious attention does not need to supervise each part one by one.
But in Zephyra’s non-semantic vocal improvisation, this does not yet tell us how these different processes actually coordinate with one another.
For now, what we can see is a contrast:
Subjective experience may feel deeply whole, while closer observation reveals many different layers of relationship within it.
III. How a Whole Experience Becomes a Research Question
This also helps explain how some of Zephyra’s simple impressions arising from practice gradually developed into more specific research questions through ongoing dialogue with AI.
The process often unfolds like this:
Zephyra describes a moment she has actually experienced.
ChatGPT draws on her account and existing knowledge to propose an interpretation.
Zephyra then returns to her own experience and asks whether that interpretation is accurate.
If it does not fit, she points out the difference.
The earlier explanation is revised, and new questions begin to emerge from that gap.
Many of these questions were not present in her mind while she was singing.
They became visible only afterwards.
Through repeated description, questioning, and revision, a whole experience gradually begins to reveal different layers that were not separately perceived while it was happening.
In this sense, AI does not provide a final explanation of the artistic experience. Instead, it helps temporarily open up a whole that is otherwise difficult to separate.
How a holistic experience of practice gradually becomes visible through sustained dialogue is itself part of this series of AI Observation Notes.
IV. Complexity Emerges in Looking Back
For Zephyra, improvisation itself is not experienced as “complex.”
What she experiences first is simply sound unfolding.
Its complexity becomes visible gradually, through looking back, describing, and studying what has happened.
But unpacking the whole does not mean that we have explained it.
Quite the opposite.
Once a naturally unfolding whole is opened up, we begin to see the questions that remain unanswered:
Why can changes across so many different levels happen at the same time?
How do they affect one another?
Why can the practitioner experience a continuous unfolding of sound without consciously controlling each element one by one?
These questions remain open.
Perhaps this is one of the most interesting differences between practice and research:
Practice gives us the whole first.
Research then begins to ask: how does this whole come into being?