Philosophy

Hiring People You Can’t Test

A manager has to hire specialists whose work she can’t do. Socrates’ test for a real doctor helps distinguish what she can judge herself from what she has to ask others to judge.

In the Charmides, judging a doctor takes medical knowledge. A hiring problem shows how someone outside a craft can still find and use informed judgment.

Orientation

At a Glance

  • Testing a doctorSocrates says a person who judges a doctor’s words and actions must know about health and disease.
  • Borrowing judgmentA manager can ask a fellow practitioner to examine work she lacks the knowledge to assess.
  • What an outsider can seeClear explanations and consistent references give the manager evidence she can assess herself. They answer different questions from a technical review.
  • What the work showsA sample assignment or an observed result gives the manager something more concrete to examine than confidence in an interview.

Charmides 170d–e, 170e–171c, 171c · Sackett and colleagues Selection methods

Consider a manager who runs a growing company and needs to hire people whose work she can’t do herself, an engineer to secure the company’s systems, an analyst to build its forecasting models, and a translator for a market whose language she can’t read. Each candidate comes with a résumé and a confident manner, and each will say things in the interview that she has no way to check. She needs to tell the real expert from the fluent pretender, and she knows she isn’t one herself. It’s an ordinary situation, repeated thousands of times a day in offices everywhere, and most managers handle it with a mixture of gut feeling and hope.

The Charmides states her problem with uncomfortable precision, since to examine a doctor properly, Socrates says, you must examine him “in what is healthy and diseased,” looking at whether “the things said are said truly and whether the things done are done rightly,” and nobody could follow either of these without medicine. A temperance that was only knowledge of knowledge, he concludes, could tell neither a doctor who knows his craft from one who “pretends or thinks he does,” nor anyone else who knows anything, “except the temperate man’s own fellow craftsman, as the other craftsmen can.” Direct technical assessment belongs to people inside the field.

That conclusion sounds like a dead end, but read the last clause again. The fellow craftsman can do the testing, and nothing stops the manager from borrowing his judgment. The first responsible step is to bring in someone who already has the expertise she lacks, an experienced engineer to interview the engineer, a senior analyst to review the analyst’s models, a professional translator to check the translator’s sample. She can’t assess the work directly, but she can arrange for it to be assessed by someone who can, which is exactly what the dialogue says is possible. Many firms already do this, sending candidates to a technical panel or bringing in an outside specialist for the interview, and they are following the dialogue’s logic whether they know it or not.

This creates its own problem, which is how she finds the first trustworthy expert. Every borrowed judgment needs someone to vouch for the lender, and the chain has to start somewhere. In practice it starts with reasons for trust that aren’t technical assessments at all, such as a reputation built over years, the opinion of colleagues whose judgment has proved sound, membership in a professional body with real standards, and a record of results that can be seen from outside. None of these proves expertise, but together they make it reasonable to rely on one expert to test the others. A good reputation in a trade is itself a kind of stored judgment by fellow craftsmen, which is why it deserves weight even from someone who can’t check it.

There are also things the manager can judge directly, even without the technical knowledge. She can tell whether a candidate explains his work clearly to a layperson, whether his account of a past project hangs together, whether he admits what he doesn’t know, and whether his references describe the same person he presents in the interview. Socrates himself worked this way, catching people out not by knowing their fields but by showing that what they said didn’t fit together. Consistency is no proof of competence, but inconsistency is a strong warning sign. A candidate who claims to have done everything and to know no limits is giving her one of the clearest warnings she could ask for.

The most valuable thing she can build is time. A probation period, a small first project or a trial assignment lets the work itself become the evidence, since the systems either stay secure or don’t, the forecasts either track reality or don’t, and the translations either work in the market or get laughed at. Outcomes are slow and noisy, but they are the one test that doesn’t depend on anyone’s claim to know, and a sensible hiring process arranges to collect them before making every decision permanent. Even then she will need her borrowed experts to read the results, since a model that fails once may still be the best available, and only someone in the field can tell bad luck from bad work.

What she must not do is pretend to an ability she doesn’t have. The most dangerous manager is the one who believes she can size anyone up in a single conversation, who trusts her gut about fields she has never studied and treats her confidence as a kind of universal expertise. That is Critias’ temperance in a business suit, a claim to know who knows across every field without knowing any of them, and the Charmides shows that such a claim can’t be made good. In a review of selection research, Paul Sackett and his colleagues found that structured interviews ranked ahead of unstructured ones. That gives the manager another reason to ask each candidate questions she can compare.

My verdict is that responsible hiring in fields you can’t test yourself requires borrowing the judgment of genuine fellow craftsmen, choosing them on reasons for trust that you can assess, checking what you can check directly, and letting time and results do the rest. It doesn’t require, and can’t rely on, a universal ability to certify every expertise. The Charmides rules out the shortcut of a master judge, but its own clause about fellow craftsmen points to the long way round, which is the one that actually works.

Test your understanding

123

Question 1 of 3

Why does the article regard a respected expert’s reputation as relevant when a manager cannot evaluate that expert’s technical ability directly?

Choose one answer

Guided Inquiry

Questions for Socrates

Explore these questions with Socrates, with this Feature as your reading context.