Showing in Design

In the AI Story

The concept emerges from Crawford's broader framework of craft practice but sharpens in the AI context because AI specifically operates through the specification structure showing resists. A specification describes what the output should do — its functional requirements, its performance parameters, its structural constraints. A well-made object does what the specification requires but also shows qualities the specification could not describe — the proportion that feels right in a way no formula predicts, the weight that sits in the hand with authority, the joint that reveals the maker's understanding of how the piece will be used.

The showing dimension is what makes craft quality resist both mass production and AI-directed design. Mass production can replicate specifications but cannot replicate the judgment that produces the qualities specifications do not capture. AI-directed design can generate outputs meeting specifications faster than human designers, but the acceleration is in the specification-meeting dimension — the dimension in which showing does not live. The chairs produced by AI-directed processes are not less functional than craft chairs; they are less capable of showing what craft chairs show, because the process that would develop the judgment producing showing has been bypassed.

The philosophical dimension connects to Crawford's broader engagement with tacit knowledge. Showing is the manifest form of tacit knowing — the qualities of a made thing that reveal the maker's embodied understanding without requiring that understanding to be propositionally articulated. A well-made chair shows the maker's judgment about how bodies sit, how wood behaves, how proportions relate, in the same way Michael Polanyi's famous example of a good face shows the recognizable person without requiring each feature to be specified in advance. The showing is holistic, integrative, and irreducible to the parts its specifications would describe.

The implication for AI-directed design is not that AI cannot produce useful objects — it obviously can. The implication is that a culture whose designed environment is increasingly produced through AI-mediated specification will progressively lose the showing dimension of its material culture. The objects will work. They will meet requirements. They will not show the qualities that distinguish a culture that cares about what it makes from a culture that merely produces what its specifications describe.

Origin

The concept is Crawford's own — an extension of his craft philosophy developed in explicit engagement with the AI transition's specification-based production model. The terminology draws on Wittgenstein's distinction between saying and showing in the Tractatus but applies it to material culture rather than linguistic philosophy.

The philosophical tradition the concept engages includes David Pye's distinction between workmanship of risk and workmanship of certainty, Christopher Alexander's attention to the quality without a name, and the broader phenomenological tradition's treatment of what shows itself in made things without being explicitly represented.

Key Ideas

Showing vs. specifying. The distinction between what a well-made thing shows to sustained attention and what its specifications can describe — a distinction AI-directed design systematically fails to preserve.

Holistic integration. Showing is a property of the integration between function, form, material, and maker's judgment — not a feature that can be added to specifications as an additional requirement.

Manifest tacit knowledge. Showing is the manifest form of the maker's tacit understanding — the qualities of a made thing that reveal embodied judgment without requiring propositional articulation.

The specification-based production model. AI-directed design accelerates production in the specification-meeting dimension but does not operate in the dimension in which showing lives.

Cultural implications. A culture whose designed environment is increasingly AI-mediated will progressively lose the showing dimension of its material culture — producing functional objects without the qualities that distinguish thoughtful making from mere production.

Debates & Critiques

The challenge to the showing concept is that it risks being unfalsifiable — any claim that something is missing from AI-produced designs can be met with the response that the perceived absence is subjective aesthetic preference rather than objective design deficit. Crawford's response is that showing is detectable through specific practices (sustained comparison, use over time, maintenance and repair experience) that reveal qualities invisible to surface inspection. The detection is empirical, not merely subjective. Whether the empirical work required to detect showing scales adequately for contemporary consumer goods remains an open question.