Mill's Handle

Mill's chapter on naming as subsidiary to induction argues that names give thought a point of attachment, while explicitly denying that names are indispensable to reasoning. Lupyan, Rakison and McClelland found in 2007 that even redundant labels helped people learn novel categories, and that a nonverbal association did not. Against that sits reification: once a proposed pattern becomes a noun, sentences begin treating it as an established object. The ledger's own architecture creates the bias, which is why inclusion here is not endorsement.

In practice

• A team that names a recurring failure can search for it, count it and argue about its boundaries, none of which it could do while the failure was only a story.
• The same name, repeated confidently for a year, has the department certain a thing exists that nobody has yet measured.

Possible explanations

The explanation is retrieval and comparison: a named pattern can be recalled, matched against a new case and argued about, where an unnamed one has to be reconstructed each time.

In its favour

Mill supplies the historical argument and experimental work on labels and categorisation supplies evidence that naming changes what people notice and remember.

Against it

It is a coinage for this ledger and a synthesis rather than a single finding, and the same property makes bad claims travel, which is the cost the principle carries with it.

Standing

A coinage for this ledger and a synthesis rather than a single finding. Mill supplies the historical anchor, Lupyan the experimental evidence on labels, Fricker the collective case and Hacking the looping caution; the predictions it generates are largely untested. Set out in full in the essay of the same name. Sailer's Law of Neologisms disputes part of it, holding that self-explanatory names are precisely the ones that fail to travel.

Subject

Science & Epistemology

Essay

How knowledge grows, and how inquiry fools itself

Read the essay on Mill’s Handle