3 min read

Proofreading Science: See What is There

Henri Poincaré famously, and correctly, admonished assistants who brought grammatical errors to his attention. Focus placed on polish tends to be at the expense of substance, a common bildungsphilister characteristic. With minds at their acme in trivia, detail polishers are largely incapable of the introspection and intuition required for discovery. That the former is selected for over the latter in university science departments is sadly evident. There is an analogue to proofreading, however, that has nothing to do with grammar and is quite necessary for science.

Proofreading requires that one sees what is there, rather than what is expected. In writing, errors of spelling and punctuation slip past the undiscerning eye. In science, the risk of error is also present. Like a proofreader, a scientist must see what is there rather than what is expected.

If a series of chemical assays has an expected result, how much deviation from the expected must there be to take notice? If noticed, what is given priority, the observation or the expectation? What level of contradiction between experiment and theory must be attained to question the theory and not the experiment? Are there incentives to the scientist for one over the other?

Many would be surprised at how crude the tools of science can be, and by the degrees of separation between them and what they purport to measure. A physical phenomenon, framed in a scientific question, cannot be studied directly. Assays and models act as handles with which to inspect and manipulate the question. The more and varied handles the better, as they work to corroborate each other. A problem arises when the assays and models are equated not to measurements, but to phenomena per se. For example, a measurement of cytokines produced during inflammation becomes a measurement of inflammation; a biomarker associated with heart disease becomes a test for heart disease. The numbers acquired from these tests become the definition of a phenomenon or disease. This tendency creates a tautology factory, where new and exciting ways to ask scientific questions mold the psyche of the field, transmogrifying into answers in and of themselves.

One obvious reason for this tendency is the incentives produced by the funding structure. All centrally planned, coercive schemes lack the ability to rationally allocate resources (1). State agencies for the funding of science must use some metric to determine the dispensation of scarce funds. Similar to Hollywood, which likes the latest and greatest special effects, the National Institutes of Health have a penchant for cutting-edge techniques. The bureaucrats that oversee funding have to cover their hides, and defend from charges of incompetence. Jargon and complicated techniques, the more incomprehensible the better, perform this function. Cutting-edge techniques and novel theoretical models are recognizable units of progress that can fill the forms of bureaucrats and politicians.

The tautology factory has another, less obvious, cause. Imperfect as they are, the scientist’s tools are all that he has. The fundamental inability to directly study a phenomenon produces anxiety. Handles provided by the tools of science direct the scientist’s inquiries, and his mind too conforms to their contours. The concrete numbers displayed by his tools soothes his anxiety. A real emotional incentive exists to substitute the map for the terrain.

Once an assay serves to confirm itself, there is a further tendency to see only confirming data. If 4 out of 5 assays confirm each other to tell a story, the 5th becomes invisible, as one wrong letter in a word is glossed over. The difference between proofreading the written word and proofreading science is that the written word represents a known concept; science attempts to discover a new one. Misspelling a word does not affect its concept. Only if a word is butchered to the point of incomprehensibility will it impair communication. Glossing over misbehaving data in the service of preserving a narrative always impedes discovery and understanding, sometimes completely.

The most valuable asset for a scientist is the beautiful mind, a creative mind capable of consideration from multiple angles. An absolute prerequisite is the ability to see precisely and only what is there. Without the bonds of honest data assessment, the beautiful mind will indulge in flights of fancy unrelated to true discovery and understanding. What’s worse, the vulgar mind will follow the incentives set by funding agents, set by politicians.


1) Mises, Ludwig von, 1990, Economic Calculation in the Socialist Commonwealth, Auburn, AL: Ludwig von Mises Institute.