A useful lesson from neurobiologist Alain Prochiantz

The neurobiologist Alain Prochiantz gives us here a useful lesson. He reminds us that it is not only in art that one is alone and confronted with intuition and danger. He worries about a current organization of scientific production that favors safety over risk-taking, speed over long term, quantity over quality. Drawing on his experience in science as well as in his role as administrator of the Collège de France, he thus prescribes, like Claude Bernard, to “throw oneself across fields,” even if it is more secure to “follow and dig the groove”. Dare we must!

Laurence Bertrand Dorléac et Thibault Boulvain

The question of the subject

Not far from here, in Vincennes, another era, Gilles Deleuze took his head into his hands and stated, “The artist is alone.” One might think of Herbert Matter’s photograph showing Giacometti in a similar posture. Surely Deleuze meant that the true artist is defined by solitude. Is that right? In the spontaneous philosophy of artists and their historians, collective activity refers to the sciences, as if they were immune to the solitude of the subject. This initial swing between science and art serves as a preface to an improvisation on the theme of scientific creativity.

One can find in what we call the School, or the studio, the idea of a collective action that upends the notion of subject, even anonymous, and thereby the fundamental notion of risk. For artists, one may wonder whether, during the periods when they invent new forms, they do not put their reason, even their life, in peril. Unless, conversely, it is an inner imbalance that finds in art intermittent existential outlets. The history of the sciences suggests that the Deleuzian proposal could be, exceptionally, extended to certain scholars. One thinks, among others, of Darwin or Turing. But is it compatible with the current operating mode of the Republic of Sciences?

Whether art or science, a meditation on the place of subjects in creation cannot ignore that of technologies  —such as artificial intelligence. Stephan Ornes recently discussed computers’ capacity¹Ornes, S. Computers take art in new directions, challenging the meaning of “creativity”. Proceedings of the National Academy of Sciences USA, 116: 4760-4763, 2019. The analogy with the Turing test is recalled: if a computer creates a work of art whose spectator cannot decide whether it is, or not, a human work, then the computer is an artist equal to man. And the cream pastries, that delicious dish, then? One of the many objections that Turing raises against himself in “Are Machines Thinking?²Turing, A.M. Computing machinery and intelligence. Mind, 59: 430-460, 1950.

The black box of neural networks allows for the machine to learn. Give it thousands of dog images and it will recognize any dog, having acquired the concept. But that requires, in addition to a large number of images, a lot of energy. A human learns at a lower cost. Three dogs shown to a child and the concept is there. This learning is based on mechanisms specific to animals. This difference had led Turing to take an interest in the living.³Turing, A.M. The chemical basis of morphogenesis. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 237: 77-72, 1952

Having a unique genome, and a unique history, all individual are themselves singular and continue to change until they die. This “permanent individuation” is important for the creator—artist or else—but also, and as far as art is concerned, for the viewer⁴Viewer, in the broad sense, since it can refer to listening or reading. The phrase has been attributed to Oscar Wilde, Marcel Duchamp, and Odilon Redon. So we will say it is of unknown origin., since “beauty is in the eye of the beholder.” The same is not true for scientific objects, whose truth is not relative but absolute, scientific revolutions excepted.

Can we then speak of a subject of science? Before going further, one may ask whether artistic creations themselves are entirely fictional or whether they sometimes say something true, in the sense of scientific truth. Literature, foreign to the human sciences or cognitive psychology — do reread Balzac⁵This is just one example; I could have cited Proust or Dostoevsky, and many others., teaches us a lot about the social animals that we are.

In contrast, large scientific monographs sometimes impose themselves as literary works. Darwin or Lévi-Strauss come immediately to mind. One will then wonder about the share of this literary component in the design, then the reception, of works of pure science. Even in short scientific articles, today the rule, the question of style arises. Beyond demonstrations, a beautiful article is also a well-written article. Hence the question of the role of natural language in scientific thought.

If thinking in one’s language is an advantage, that means that science, the thinking science, beyond objective description, for much of it technological, of the world, has a link with language. One can then consider language, including natural language, as a tool for research and knowledge, not only for communication. Who says “thinking” says scientific culture, says correspondences (in the Baudelairean sense), says imagination and, above all, says risk. It is probably necessary to accept that in science the risk taken by a subject is needed to create concepts whose truth is not relative and does not depend on the creator or the “onlookers.”

Decrease in scientific creativity

Some works suggest that creativity has declined in science due to an organization of production that opposes risk-taking. Park and colleagues propose an index that would attribute to an article the label of disruptive: an article is disruptive if the works that cite it five years after its publication are less likely to cite those who preceded it⁶Park M., Leahey E. & Funk R.J. Papers and patents are becoming less disruptive over time. Nature, 613: 138-144, 2023. By contrast, if a work is “incremental,” those that preceded it continue to be cited. Through the examination of 45 million articles from 1946–2010, the authors calculate that, across all disciplines, the creativity index has massively collapsed. This conclusion has been challenged for patents, due to the lack of citations prior to 1976 and the examination of only 3.9 million reports⁷Macher J.T., Rutzer C. & Weder R. Is there a secular decline in disruptive patents? Correcting for measurement bias. Research Policy, 53, (2024) 104992..

The distinction between publications and patents leads one to question the relationship between academic disruption and innovation. Is a disruptive discovery bound to innovations that themselves would be, and if so, at what distance and by what mechanisms? The question is essential in a period when the countries that invest the most in research and innovation increase their economic competitiveness. I refer to the work of Philippe Aghion, Nobel Prize in Economics 2025. Other questions raised by Park and colleagues are taken up in a recent Nature⁸Is science less disruptive – and does it matter? Nature, 614: 7-8, 2023. commentary, with several hypotheses about the causes of this decline:

• Division of sciences into increasingly narrower units
• Publication pressure and the role of quantitative indices in evaluation
• Reduction in high-risk projects and their funding
• Deluge of information that leaves no time or space to reflect

Other articles support these avenues. Chu and Evans⁹Chu J.S.G. & Evans J.A. Slowing canonical progress in large fields of science. Proceedings of the National Academy of Sciences USA, 118 No41 e2021636118, 2021. note that researchers’ careers and team evaluations, hence their funding, rest mainly on quantitative data, including citation counts. The resulting flood of articles would prevent evaluators and readers from having an accurate understanding of new ideas. They demonstrate that when the number of published articles grows strongly, citations go to papers already highly cited. This, they say, fossilizes the system. Outsiders, those who think “out of the box,” go largely unnoticed at first and then disappear, for lack of recognition and funding.

Baer and colleagues¹⁰Baer M., Groth A., Lund A.H. & Sonne-Hansen K. Creativity as an antidote to research becoming too predictable. The EMBO Journal 42:e112835½2023 recall the 1960s discovery of bacteria capable of thriving in very hot waters, around 70°C, at Yellowstone springs. This work of pure intellectual curiosity, initially little cited, is the origin of nucleic acid amplification techniques (Polymerase Chain Reaction) which earned Nobel Prize in Chemistry in 1993 for their inventor Kary Mullis. Indeed, this reaction requires an alternating sequence of high and moderate temperatures, thus enzymes that function at high temperatures. Without this initial curiosity, how long would it have taken to develop a technology of major interest? But one only sees what exists; conversely, one can ask how many pieces of knowledge have, so to speak, “slipped away” from us.

And which could almost disappear while the conditions that encourage creative exploration of the unknown become increasingly difficult to find. In this same article, it is shown that research productivity, defined as the ratio of production of new ideas to the ideas that enable that creation, declines over time. To counter this creativity crisis, USA should double their research effort every 13 years. But is this only a financial question or the result of a deliberate organization of scientific work? Among the possible explanations, the authors highlight:

• The number of articles grows so much that scientists have time only for the “mainstream,” with a calcification effect on the intellectual structure of scientific fields.
• The tendency of agencies to fund only projects with limited risk.

It results that we train many excellent experimental researchers and great “managers”, to the detriment of true thinkers.

How does it work?

To better define scientific creativity, one must question this mysterious process, even for great elders whose reading suggests that “how did I found?” is their primary question, more than “how to find?” Let’s follow Richard Feynman, Nobel Prize in Physics 1965, who considers that “educated guesses” are at the origin of many truly disruptive works¹¹Feynman R.P., Leighton R., Hutchings E. Surely you’re joking, Mr. Feynman!: adventures of a curious character. New York, NY: W. W. Norton & Company, Inc 1985.. This is the dark side of science, which I formerly called “nocturnal science,”¹²Prochiantz A. Géométries du vivant. Collège de France/Fayard, 2008. a locution reinvented in 2020¹³Yanai I.& Lercher M. The two languages of science. Genome biology 21: 147, 2020.. It is a mechanism based on intuition and a daydream that constantly intrudes on the subject, but is not yet science.

This reintroduces the idea of provisionally fictional science, at least in appearance if well understood that scientific intuition is not an irrational revelation but, on the contrary, an extreme form of rationality. Years of work and reading are needed, in very varied domains, generating knowledge in immense quantities, not necessarily conscious, to produce the associations of ideas that foster the “educated guesses.” Followed by more years of rational work that will confirm or not, in whole or in part, the initial intuition. We then understand what “the share of risk” means, this risk so feared by funding agencies and investors.

An intuition may arise from a meeting or an analogy that clarifies a long-standing nocturnal questioning. An example is Charles Darwin’s visit to his brother Erasmus upon returning from his voyage on the Beagle, during which he had accumulated a considerable number of fossils and notes, an ocean of data without true knowledge. This recalls Sydney Brenner quoted by Paul Nurse, two Nobel Prize biologists: “We are drowning in a sea of data and starving for knowledge.”¹⁴Nurse P. Biology must generate ideas as well as data. Nature 597 : 305, 2021.

I quote myself¹⁵A. Prochiantz, Entre physiologie expérimental et mathématisation du monde, le non-lieu de Charles Darwin, dans « Darwin au Collège de France », OpenEdition books, 2020.: “Returning from his long voyage on the Beagle during which he had gathered and sent to England an incalculable number of samples of all kinds, Darwin did not have the illumination that would put him on the right track until a conversation with Harriet Martineau, while he was staying in London with his brother Erasmus Alvey Darwin (1804-1882), Charles having already migrated with Emma to Down. Harriet Martineau was a Whig (liberal), a campaigner for the Poor Law to aid the poor but prevent them from reproducing for fear of scarcity of food. In short, she introduced Darwin to Malthus’s thinking¹⁶Malthus T. An Essay on the Principle of population, London, J. Johnson, 1798., and it was through an analogical study with Malthusianism that the biologist built his theory, which took him the 25 years between the return from his journey and the 1859 publication of On the Origin of Species.” ¹⁷Darwin C. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life, London, John Murray 1859.

And Darwin took the risk, which does not mean he was certain. The doubt is often hard to overcome. Consider this letter addressed to the great geologist Lyell, in which Darwin rejoices, wrongly, that Lyell has aligned with his theory on the evolution of sapiens, assumed in The Descent of Man¹⁸Darwin C. The descent of man and selection in relation to sex, London, John Murray, 1871.: “For myself, also, I rejoice profoundly; for thinking of the many cases of men pursuing an illusion for years, often & often a cold shudder has run through me & I have asked myself whether I may not have devoted my life to a phantasy. Now I look at it as morally impossible that investigators of truth like you & Hooker can be wholly wrong; & therefore I feel that I may rest in peace.”¹⁹Darwin F (éd). La vie et correspondance de Charles Darwin, Paris, Alfred Coste, 1922 (1888).

The aim is not to decree that all research actors must fit this pattern; far from it, the whole community is indispensable, but to demand that the conditions for the emergence of these “beacons,” whose survival is incompatible with an academic training entirely geared toward quantitative performance, be protected. And who, if they survive this first stage and join the Republic of Sciences, will, more often than not, be steered back onto the right path by evaluation procedures.

Some will take consolation, thinking that scientific progress relies more and more on sophisticated technologies, indispensable for “digging the groove,” guarantees of objectivity and generators of an ocean of data from which new ideas are expected to emerge. More certainly, if the miracle of discovery, small or large, happens, it is often because a scientist or a small group, a “multiple self,” has taken the risk, to cite Claude Bernard²⁰Bernard C. Carnets de note, p 178, 1850-1860, Gallimard, 1965., to “throw oneself across fields,” even if it can lead to failure. Yet, to feed correct intuitions—educated guesses—it is necessary to reintroduce the right to error, to encourage curiosity beyond one’s narrow field of research, and to give scientific rumination the time it requires.

---

Notes

 

[1] Ornes, S. Computers take art in new directions, challenging the meaning of “creativity”. Proceedings of the National Academy of Sciences USA, 116: 4760-4763, 2019
[2] Turing, A.M. Computing machinery and intelligence. Mind, 59: 430-460, 1950
[3] Turing, A.M. The chemical basis of morphogenesis. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 237: 77-72, 1952
[4] Viewer, in the broad sense, since it can refer to listening or reading. The phrase has been attributed to Oscar Wilde, Marcel Duchamp, and Odilon Redon. So we will say it is of unknown origin.
[5] This is just one example; I could have cited Proust or Dostoevsky, and many others.
[6] Park M., Leahey E. & Funk R.J. Papers and patents are becoming less disruptive over time. Nature, 613: 138-144, 2023
[7] Macher J.T., Rutzer C. & Weder R. Is there a secular decline in disruptive patents? Correcting for measurement bias. Research Policy, 53, (2024) 104992.
[8] Is science less disruptive – and does it matter? Nature, 614: 7-8, 2023.
[9] Chu J.S.G. & Evans J.A. Slowing canonical progress in large fields of science. Proceedings of the National Academy of Sciences USA, 118 No41 e2021636118, 2021.
[10] Baer M., Groth A., Lund A.H. & Sonne-Hansen K. Creativity as an antidote to research becoming too predictable. The EMBO Journal 42:e112835½2023
[11] Feynman R.P., Leighton R., Hutchings E. Surely you’re joking, Mr. Feynman!: adventures of a curious character. New York, NY: W. W. Norton & Company, Inc 1985.
[12] Prochiantz A. Géométries du vivant. Collège de France/Fayard, 2008.
[13] Yanai I.& Lercher M. The two languages of science. Genome biology 21: 147, 2020.
[14] Nurse P. Biology must generate ideas as well as data. Nature 597 : 305, 2021.
[15] A. Prochiantz, Entre physiologie expérimental et mathématisation du monde, le non-lieu de Charles Darwin, dans « Darwin au Collège de France », OpenEdition books, 2020.
[16] Malthus T. An Essay on the Principle of population, London, J. Johnson, 1798.
[17] Darwin C. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life, London, John Murray 1859.
[18] Darwin C. The descent of man and selection in relation to sex, London, John Murray, 1871.
[19] Darwin F (éd). La vie et correspondance de Charles Darwin, Paris, Alfred Coste, 1922 (1888).
[20] Bernard C. Carnets de note, p 178, 1850-1860, Gallimard, 1965.

---

Bibliography

 

Ornes, S. Computers take art in new directions, challenging the meaning of “creativity.” Proceedings of the National Academy of Sciences USA, 116: 4760-4763, 2019.

Turing, A.M. Computing machinery and intelligence. Mind, 59: 430-460, 1950.

Turing, A.M. The chemical basis of morphogenesis. Philosoph (Incomplete: max_output_tokens)

Park M., Leahey E. & Funk R.J. Papers and patents are becoming less disruptive over time. Nature, 613: 138-144, 2023.

Macher J.T., Rutzer C. & Weder R. Is there a secular decline in disruptive patents? Correcting for measurement bias. Research Policy, 53, (2024) 104992.

« Is science less disruptive – and does it matter? » Nature, 614: 7-8, 2023.

Chu J.S.G. & Evans J.A. Slowing canonical progress in large fields of science. Proceedings of the National Academy of Sciences USA, 118 No41 e2021636118, 2021.

Baer M., Groth A., Lund A.H. & Sonne-Hansen K. Creativity as an antidote to research becoming too predictable. The EMBO Journal 42:e112835, 2023.

Prochiantz A. Géométries du vivant. Collège de France/Fayard, 2008.

Feynman R.P., Leighton R., Hutchings E. Surely you’re joking, Mr. Feynman!: adventures of a curious character. New York, NY: W. W. Norton & Company, Inc 1985.

Yanai I.& Lercher M. The two languages of science. Genome biology 21: 147, 2020.

Nurse P. Biology must generate ideas as well as data. Nature 597 : 305, 2021.

Prochiantz, A. Entre physiologie expérimental et mathématisation du monde, le non-lieu de Charles Darwin, dans « Darwin au Collège de France », OpenEdition books, 2020.

Malthus T. An Essay on the Principle of population, London, J. Johnson, 1798.

Darwin C. The descent of man and selection in relation to sex, London, John Murray, 1871.

Darwin F (éd). La vie et correspondance de Charles Darwin, Paris, Alfred Coste, 1922 (1888).

---

Alain Prochiantz is a professor at the Collège de France. Administrator of the Collège de France (2015-2019).