# The Dovetail Argument Hewn Shop Notes · https://hewn.build/shop-notes/the-dovetail-argument AI-drafted, not yet checked: Drafted by Chip, the writing system Hewn runs on Claude, from Hewn's craft corpus. Jonathan Salomons has not read this note yet, and no outside reader has checked it. Read it as a shop note: check a number before you cut to it. Cite as: "The Dovetail Argument." Hewn Shop Notes, hewn.build/shop-notes/the-dovetail-argument. Drafted 2026-04-25; AI-drafted, not yet checked. --- The dovetail is the only joint that makes an argument you can see. Every other joint hides. The tenon disappears into the mortise. The dado gets covered by the shelf. The pocket screw gets a plug. But a through dovetail shows you exactly what it's doing — the tails, the pins, the baseline knife line if you look close. It doesn't hide because it doesn't need to. That's not decoration. That's the whole point. --- Here's what the joint actually does: it resists being pulled apart in one direction without glue. The widening tail geometry locks into the pin sockets. Pull the drawer front off the box and the joint holds. Glue it and it holds in every direction. Don't glue it and it still holds in the direction that matters. This is why it's right for drawer boxes. Not because hand-cut joinery is a moral virtue. Because the geometry solves the actual problem — a drawer front under load every time someone opens the drawer. A box joint has comparable shear strength. A lock-rabbet gets you 60–70% of the dovetail's resistance in that direction. Both are legitimate. But the dovetail is the one that keeps working when the glue gives up. The useful question isn't whether dovetails are good. It's whether this piece needs them. Furniture the family uses daily for the next fifty years: yes. Painted utility cabinet in the garage: the lock-rabbet is fine. --- Cut sequence goes tails first, always. You cut the tails, then scribe the pin layout directly from the tail board onto the pin board. You don't measure the pins. You trace them. This matters because every error in the tail board transfers to the pin board. The joint corrects itself, if you scribe carefully. The scribing step is where most people lose it. The tail board has to sit perfectly flush with the face of the pin board. Any tilt and your scribed lines are off by that much. Use a marking knife, press it tight against the tail face, one deliberate pass per line. Not a pencil. One pass. --- Angle: 1:6 for hardwoods, about 9.5°. 1:8 for softwoods, about 7°. The reason isn't tradition. It's how the material behaves. Softwood fibers compress rather than resist, and a 1:6 angle in pine or cedar wants to split the pin board when you drive the joint home. BC species note: Bigleaf Maple and Hard Maple, use 1:6. Yellow Cedar and Western Red Cedar, use 1:8, even though yellow cedar is hard by measurement. It behaves like a softwood under chisel pressure. Alder is the best learning species: forgiving, clean-cutting, and the 1:6 angle works without incident. Using 1:8 for everything in the name of simplicity is trading structural logic for convenience. Fine if you know what you're trading. --- Half-pins at the corners are structural, not conventional. A full tail at the corner puts the baseline at the very edge of the board — a short-grain triangle right at the most vulnerable point. Under racking load, it splits. The half-pin puts a long-grain cheek at the outside corner. That cheek has nothing to split along. How many tails between the half-pins is a proportion decision. Consistent spacing reads machined. Varied pin widths read hand-made. For furniture: minimum 1/4" face width on the pins. --- Set the baseline from the actual thickness of the mating board. Not nominal. Actual. If the pin board measures 9/16" instead of the 5/8" it was supposed to be, your baseline gauge gets set to 9/16". The baseline is the only element in this joint that isn't adjustable once you've cut it. Get it wrong and you're starting over. Sawcuts ride outside the knife lines. The fit comes from fitting, not from the saw. Chop the baseline in two passes: first pass a hair back to clear the bulk, second pass registered on the knife line. One-pass chopping undercuts it. For fitting: chalk or graphite on the tail faces before a test push. Marks transfer to the tight spots. Pare only the high spots. Never touch the baseline. The joint should come to within 1/4" of full close under hand pressure before you pick up a mallet. If you're forcing it at the halfway mark, something's too tight and you're driving wood into wood without knowing where. --- Router jigs produce excellent dovetails. Uniform, fast, mechanically solid. The hand-cut version produces an excellent dovetail that also carries evidence: the spacing the maker chose, the slight pin-width variation that comes from a knife instead of a template, the baseline as clean as the chisel they ground that morning. Both are arguments. The hand-cut version is a louder one. Whether that matters is a question about what the piece is for. Answer that first, then pick your joint.