We control conditions — but the system responds through interactions.
- We describe fermentation as a process.
- We define inputs, set parameters, and expect outputs.
- That description is convenient.
- It is also incomplete.
What We Assume
- A process implies stability.
- You set temperature, pH, feed rate.
- You monitor dissolved oxygen, biomass, metabolites.
- The system is expected to respond in a controlled, predictable way.
- If something deviates, you adjust a parameter.
- This framing suggests that the system is driven by inputs.
- That behavior follows control.
What Actually Happens
- Fermentation does not behave like a linear process.
- It behaves as a network of interactions — between cells, environment, and time.
- Microbial populations shift.
- Metabolic states change.
- Local conditions diverge from bulk measurements.
- These changes are not always visible as they occur.
- They accumulate, interact, and then appear.
- What looks stable at the process level often hides continuous change underneath.
The Illusion of Control
- We control parameters.
- We do not control the system.
- Adjusting pH or oxygen does not directly control cellular behavior.
- It influences conditions that cells respond to — in ways that depend on context.
- Two systems with the same setpoints can evolve differently.
- The difference is not noise.
- It is interaction.
Where Measurement Falls Short
- Most process measurements are aggregated.
- They reflect averages across the system — not local dynamics.
- Dissolved oxygen is reported as a single value.
- pH is treated as uniform.
- Biomass is measured as a bulk quantity.
- But the system is not uniform.
- Gradients form.
- Microenvironments emerge.
Cells experience conditions that are not captured by the measurements we rely on.
We observe a simplified version of the system — and treat it as complete.
Timing Is Structural
In fermentation, timing is not just a measurement issue — it is part of the system itself.
Changes in metabolic activity propagate through the system before they become visible in aggregated signals.
What we detect is not the shift —
but its downstream effect.
Stability Is an Output
What we call a stable process is often a stabilized outcome.
The system has already resolved its internal dynamics into something that appears consistent.
But that stability can mask sensitivity.
Small shifts — in population balance, substrate availability, or local conditions — can move the system toward a different state long before it becomes visible.
By the time instability is detected, the system has already transitioned.
Why This Matters
If fermentation is treated as a process, we focus on control.
If it is understood as an interaction, we focus on behavior.
Control assumes predictability.
Interaction requires observation.
The Misalignment
We model fermentation as stable and controllable, but it behaves as a shifting, coupled system.
This misalignment affects everything downstream — how we measure, how we interpret what we see, and how we respond to change.
We are applying a process framework to a system that does not behave like one.
What This Leaves Us With
As long as fermentation is treated as a process, we will continue to rely on parameters and aggregated signals that describe outcomes, not dynamics.
Observing this system requires resolving behavior where it forms — not where it stabilizes.
Until then, we will keep adjusting conditions —
and miss the level where the system is actually defined.
— Pegah Farr