Sweller’s cognitive load theory: the capacity limit and its consequences

Sweller’s (1988) cognitive load theory established the foundational account: working memory — the attentional system through which all deliberate thinking occurs — has a limited capacity that, when exceeded, produces measurable degradation in the quality of the cognitive work being performed. The limit is not a threshold that, once reached, produces total cognitive failure; it is a resource constraint that, as it is approached and exceeded, progressively reduces the quality of every task that draws on it.

Cowan’s (2001) working memory research refined the capacity estimate to approximately 4 ± 1 chunks of meaningful information — a tighter limit than Miller’s original 7 ± 2, and one that better accounts for the chunks’ semantic density. The entrepreneurial cognitive load problem is structural rather than occasional: the entrepreneur simultaneously carries open investor conversations, product development decisions, team management considerations, customer relationship questions, financial planning uncertainties, and strategic options. Each of these maintains an active representation in working memory independently of whether it is being actively addressed. The aggregate load typically and chronically exceeds the capacity limit — which means that every deliberate cognitive task the entrepreneur performs is being performed from a depleted baseline.

The Zeigarnik mechanism: unfinished tasks as persistent working memory occupants

Zeigarnik’s (1927) foundational research documented the mechanism through which uncompleted tasks maintain their grip on working memory: incomplete tasks create active cognitive representations — open loops — that persist until the task is either completed or explicitly abandoned. The practical consequence for the entrepreneur managing multiple simultaneous open decisions is that the working memory is never fully available for any single task; it is simultaneously maintaining the representations of all unresolved tasks, regardless of whether those tasks are relevant to what is currently being worked on.

Levitin’s (2014) synthesis extended this with the finding that the brain cannot effectively distinguish between a task it is actively working on and a task it has merely registered as unresolved. Both consume working memory bandwidth at approximately equivalent rates. This is the explanation for the phenomenon of sitting down to work on a specific high-priority problem while simultaneously being distracted by the awareness of every other unresolved problem — the distracting awareness is not a motivational failure; it is the Zeigarnik representations of all open loops competing for the working memory that the primary task requires.

The implication is direct: the most effective cognitive load reduction intervention available is externalisation — capturing open loops in a trusted external system rather than attempting to hold them in working memory. When the external record is trusted (when the entrepreneur genuinely believes that the recorded task will be attended to at the appropriate time), the Zeigarnik loop is closed: the working memory representation is released because the task no longer requires active maintenance to avoid being forgotten. The David Allen Getting Things Done system is, at its cognitive substrate, a Zeigarnik effect management architecture — the trusted capture system’s primary function is closing the open-loop representations that are consuming working memory.

What genuinely reduces cognitive load and what merely feels like it does

The distinction between what genuinely reduces cognitive load and what produces the subjective experience of productivity without the underlying load reduction is commercially consequential.

Externalisation genuinely reduces load: the open loop in the trusted system is no longer occupying working memory. Environmental distraction elimination genuinely reduces load: the Gloria Mark (2008) interruption research documented that each interruption requires approximately 23 minutes to full reorientation — the extraneous load cost of constant notifications and open-plan environments is measured in the degraded quality of every task performed in these conditions, even when the interruptions feel manageable.

Multitasking does not reduce cognitive load; it multiplies it. The subjective experience of parallel processing is, neurologically, rapid task-switching with a transition cost at each switch — the working memory must reload the context of each task at every switch, consuming the capacity that genuine parallel processing would require. The entrepreneur who manages email while in a strategy conversation is not efficiently using time; they are performing both tasks from a fraction of the working memory that each task requires.

The smartphone presence finding is the most counterintuitive extraneous load evidence: research consistently documents that the mere presence of a smartphone on a desk reduces available cognitive capacity even when the phone is face-down and silent. The awareness that a notification might arrive creates a monitoring process — a Zeigarnik representation of the potential notification — that occupies working memory independently of whether the notification arrives. The device does not need to be in use to impose an extraneous load cost; its presence is sufficient.

The morning brain dump as the most accessible load reduction practice

The morning brain dump — writing down all open tasks, decisions, and concerns at the start of the working day — is the most widely used and most directly research-supported cognitive load reduction practice in entrepreneurial contexts. Its mechanism is the Zeigarnik effect: writing the open loops closes their working memory representations by creating the external record that replaces the internal maintenance requirement. The working memory that was simultaneously holding the investor conversation, the team hire, the product decision, and the financial question is partially released when these representations are externally captured in a trusted system.

The practice’s effectiveness is not in producing the plan — it is in releasing the cognitive resource that the unrecorded representations were consuming, making that resource available for the depth of thinking that the highest-priority work requires.

Books worth reading on this

Four Thousand Weeks by Oliver Burkeman. Burkeman’s account of the finitude of attention and time — and specifically his argument that the productivity strategies designed to do more are frequently adding to the cognitive load problem rather than solving it, because they increase the volume of open commitments rather than reducing the ones that are already open — provides the most philosophically grounded available complement to the Sweller and Zeigarnik research. His specific account of why the entrepreneurial impulse to optimise attention and time frequently produces more cognitive load rather than less maps directly onto the distinction between genuine and apparent load reduction this article describes.

If the dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Sweller, J. (1988), Cognitive Load During Problem Solving: Effects on Learning, Cognitive Science, 12(2), 257–285. Cowan, N. (2001), The Magical Number 4 in Short-Term Memory, Behavioral and Brain Sciences, 24(1), 87–114. Zeigarnik, B. (1927), On Finished and Unfinished Tasks, in Ellis, W. (Ed.), A Sourcebook of Gestalt Psychology, Harcourt. Miller, G.A. (1956), The Magical Number Seven, Plus or Minus Two, Psychological Review, 63(2), 81–97. Mark, G. et al. (2008), The Cost of Interrupted Work: More Speed and Stress, CHI 2008 Proceedings. Levitin, D.J. (2014), The Organized Mind, Dutton. Burkeman, O. (2021), Four Thousand Weeks, Farrar, Straus and Giroux.