Meshwork (Ingold)

In the AI Story

The network model has colonized twenty-first-century thought to a degree that makes alternatives difficult to perceive. Social networks, neural networks, computer networks, semantic networks — the metaphor has become the water we swim in. Intelligence, consciousness, creativity are all reconceived as emergent properties of networks. AI extends this by adding powerful new nodes to the network of human cognition. The analytic framework is powerful precisely because it abstracts away the messy, particular, lived quality of relationships into a clean topology of vertices and edges.

Ingold's meshwork refuses this abstraction. The lines that matter are lines of growth — the ancestral line of a human life, the line of a plant's growth through soil, the line of a conversation unfolding in time, the line of a practice accumulating through years. These lines are not abstract connections between pre-existing entities; they are the primary reality, and the entities we call people, species, communities are temporary condensations where lines converge. The difference is not merely terminological. A network has a static topology that can be mapped. A meshwork is always in motion; it cannot be mapped because mapping requires freezing it, and frozen, it is no longer itself.

Applied to AI, the distinction produces different diagnoses. The network question is: how does the new node affect the existing nodes? Does AI enhance, threaten, or merely add to human cognitive capacity? The meshwork question is: how does the new thread change the pattern of the whole weave? Which threads are being thickened, and which are being thinned? This is a different question, and it produces different answers. The network question can be answered in terms of productivity and capability. The meshwork question requires attending to the texture of practice — the quality of attention, the density of relational engagement, the richness of the tangle in which human life unfolds.

The diagnostic sharpens when the meshwork is understood as having a finite carrying capacity. A weave can be thickened in some regions only by thinning others. A new thread that interweaves quickly and densely — AI fits this description — does not simply enrich the weave; it redirects the pattern, pulling attention, time, and engagement away from the slower, thinner threads of embodied skill, local knowledge, and sensory engagement with materials. The framework knitters were threads in a meshwork that the power loom overwhelmed. Their skills did not disappear — they were pushed to the margins of a weave that reorganized around the dominant new thread. The warning Ingold's meshwork offers is that the same structural dynamic is now operating at the scale of cognitive work.

Origin

The concept is developed most fully in Lines: A Brief History (2007) and extended in Being Alive (2011). Ingold draws on a wide range of sources — Deleuze and Guattari's 'rhizome,' Henri Bergson's durée, phenomenological accounts of lived time — but the ethnographic grounding comes from his fieldwork on lines in human experience: the lines of ancestry traced in Skolt Sámi genealogy, the lines of movement traced in reindeer migration, the lines of growth traced in plants, the lines of making traced in weaving and writing.

The critique of network thinking is indebted to Bruno Latour's actor-network theory — which Ingold admires — while departing from it on the crucial question of whether the unit of analysis is the actant-node or the flowing line. Ingold's meshwork is closer to what Deleuze and Guattari called a rhizome, but grounded in ethnography rather than philosophy of difference.

Key Ideas

Lines, not nodes. The meshwork takes lines of movement and growth as primary, treating what look like nodes as temporary convergences.

Knots, not entities. A person, community, or practice is a knot where many lines cross and continue, not a bounded thing with internal properties.

Pattern, not topology. The question of what the weave is doing cannot be answered by mapping its connections; it requires attending to the dynamic pattern of thickening and thinning.

Threads compete for the weave's carrying capacity. A new thread that dominates the meshwork does not simply add capability — it reorganizes the pattern, pulling resources from older threads.

AI as a thick fast thread. The specific diagnostic concern is that AI interweaves at a speed and breadth that overwhelms the slower threads of embodied skill and material engagement.

Debates & Critiques

Critics have argued that the meshwork is simply a network by another name — that Ingold's refusal of node-talk is stylistic rather than substantive. Defenders argue that the temporal and developmental dimensions of the meshwork — lines that grow rather than connect — capture something that graph theory structurally cannot represent. The question for AI analysis is whether the meshwork's concept of pattern-degradation picks out a real phenomenon that the network's concept of node-density cannot.