The Gurindji Karungkarni Walk: Why Australian Elders Crossed Country Without Destinations
In the Victoria River District of Australia's Northern Territory, Gurindji elders undertook karungkarni—traversal walks across country that had no predetermined endpoint or duration. Unlike the better-known walkabout tradition often associated with adolescent rites of passage, karungkarni was practiced by experienced adults who would leave camp and walk, sometimes for weeks, guided not by goals but by what anthropologist Deborah Bird Rose called "responsive attention to what country was saying." The walker would move until the land itself indicated completion—through a specific bird's call, a change in water flow, or simply a feeling of rightness that the elder recognized through decades of attunement.
This practice contradicts our modern obsession with outcome-based planning. We've built entire productivity systems around defining objectives before taking action: OKRs, SMART goals, strategic roadmaps. Yet karungkarni suggests something radical—that there are forms of valuable work that cannot and should not begin with a destination.
The Gurindji understanding rested on what they called jintaku—a knowledge that emerges only through being present to the journey itself. An elder undertaking karungkarni wasn't gathering data to analyze later or searching for something lost. They were entering into a dialogue with country where both parties had agency. The walk would reveal what needed attention: perhaps a water source requiring ceremony, perhaps a story-site that needed visiting, perhaps nothing that could be articulated in language at all. The value lay not in acquiring predetermined information but in maintaining relational knowledge—understanding that only stays alive through continuous engagement.
This stands in sharp contrast to our extractive approach to exploration. When modern professionals "explore" a new field, investigate a problem, or conduct research, we typically define success criteria upfront. We explore to get somewhere. But karungkarni shows us a different model: movement that creates knowledge precisely because it refuses to constrain what can be discovered.
Consider how this applies to the work problem of investigating complex system failures. When production systems break, our standard approach is to assemble a team with a clear mandate: find the root cause, document it, prevent recurrence. We treat investigation as retrieval—the answer exists, buried in logs and metrics, waiting to be extracted. But anyone who has debugged truly complex systems knows this is fiction. The most important insights often emerge sideways: you're tracing a memory leak and discover an architectural assumption that's been generating technical debt for years. You're investigating latency spikes and realize your monitoring strategy has been asking the wrong questions entirely.
The karungkarni principle suggests that investigation should sometimes begin without a target. Set aside time to walk through your system—not to find something specific, but to be attentive to what the system is saying. Look at services you didn't build. Read code no one has touched in years. Follow dependency chains without purpose. This isn't undirected wandering; it's highly skilled attention without predetermined filters.
The Gurindji elder knew hundreds of indicators that non-initiates couldn't perceive. Their "aimless" walk was actually the opposite—it was aim-full in all directions simultaneously, receptive to significance wherever it appeared. Similarly, a senior engineer walking through a codebase without agenda brings years of pattern recognition, architectural intuition, and contextual knowledge. They're not being inefficient; they're allowing the system to surface what goal-directed searching would miss.
The walk ends when country indicates completion. In practice, this might mean: your investigation session ends when you've found something that shifts your understanding of the system's actual (not documented) behavior, or when you've identified a question you didn't know to ask, or when you recognize a pattern that connects seemingly unrelated issues.
Practice: The Investigation Walk
This week, schedule two hours with your most complex system or problem space. Write no investigation plan. Set no success criteria. Walk through it—code, logs, documentation, user behavior—following whatever threads draw your attention. Your only rule: notice when you're forcing direction rather than responding to what you're seeing. At the end, document not solutions but shifts in understanding. What does your system say when you stop telling it what you're looking for?
This practice contradicts our modern obsession with outcome-based planning. We've built entire productivity systems around defining objectives before taking action: OKRs, SMART goals, strategic roadmaps. Yet karungkarni suggests something radical—that there are forms of valuable work that cannot and should not begin with a destination.
The Gurindji understanding rested on what they called jintaku—a knowledge that emerges only through being present to the journey itself. An elder undertaking karungkarni wasn't gathering data to analyze later or searching for something lost. They were entering into a dialogue with country where both parties had agency. The walk would reveal what needed attention: perhaps a water source requiring ceremony, perhaps a story-site that needed visiting, perhaps nothing that could be articulated in language at all. The value lay not in acquiring predetermined information but in maintaining relational knowledge—understanding that only stays alive through continuous engagement.
This stands in sharp contrast to our extractive approach to exploration. When modern professionals "explore" a new field, investigate a problem, or conduct research, we typically define success criteria upfront. We explore to get somewhere. But karungkarni shows us a different model: movement that creates knowledge precisely because it refuses to constrain what can be discovered.
Consider how this applies to the work problem of investigating complex system failures. When production systems break, our standard approach is to assemble a team with a clear mandate: find the root cause, document it, prevent recurrence. We treat investigation as retrieval—the answer exists, buried in logs and metrics, waiting to be extracted. But anyone who has debugged truly complex systems knows this is fiction. The most important insights often emerge sideways: you're tracing a memory leak and discover an architectural assumption that's been generating technical debt for years. You're investigating latency spikes and realize your monitoring strategy has been asking the wrong questions entirely.
The karungkarni principle suggests that investigation should sometimes begin without a target. Set aside time to walk through your system—not to find something specific, but to be attentive to what the system is saying. Look at services you didn't build. Read code no one has touched in years. Follow dependency chains without purpose. This isn't undirected wandering; it's highly skilled attention without predetermined filters.
The Gurindji elder knew hundreds of indicators that non-initiates couldn't perceive. Their "aimless" walk was actually the opposite—it was aim-full in all directions simultaneously, receptive to significance wherever it appeared. Similarly, a senior engineer walking through a codebase without agenda brings years of pattern recognition, architectural intuition, and contextual knowledge. They're not being inefficient; they're allowing the system to surface what goal-directed searching would miss.
The walk ends when country indicates completion. In practice, this might mean: your investigation session ends when you've found something that shifts your understanding of the system's actual (not documented) behavior, or when you've identified a question you didn't know to ask, or when you recognize a pattern that connects seemingly unrelated issues.
Practice: The Investigation Walk
This week, schedule two hours with your most complex system or problem space. Write no investigation plan. Set no success criteria. Walk through it—code, logs, documentation, user behavior—following whatever threads draw your attention. Your only rule: notice when you're forcing direction rather than responding to what you're seeing. At the end, document not solutions but shifts in understanding. What does your system say when you stop telling it what you're looking for?