A century of Fisher. Analysis of Variance and studies on heredity, from the publication of ‘The correlation between relatives’… to today

Multidisciplinary workshop for a critical analysis of a founding text

12-13 September, 2019

Ecole Normale Supérieure, 45 rue d’Ulm, 75005, Paris

The workshop will take place 45 rue d’Ulm

- in the Amphi Galois/Rataud on Thursday 12 September;

- in the Salle Dussane on Friday 13 September.

Access to 45 rue d’Ulm requires an identity check. Only persons enrolled will be allowed to enter.

Participation in the workshop is free but enrolment is obligatory.

Enrolment can be done from the workshop’s web site.

In 1918, Ronald Fisher published an article entitled ‘The correlation between relatives on the supposition of Mendelian inheritance’ in the Transactions of the Royal Society of Edinburgh (52:499-533). In spite of a title that may seem enigmatic today, the work he presented has had a remarkable history. One hundred years after its publication, this article is still frequently cited, and the number of annual citations continues to increase [1]. The innovations it contains continue to play an important role in numerous areas, first of which are statistics and genetics – evolutionary, plant, animal and human.

Indeed, two major innovations are generally credited to this text:

1. It bases the analysis of the causes of variability of a phenomenon on a breaking down of the variance.

2. It attributes variance of a continuous characteristic (in Fisher’s text this is human height) to the individually weak effects of a large number of ‘”Mendelian factors” (the so-called “polygenic” model).

The history of this article is also due to the place the historiography of genetics reserved for it. This was the action that led to the end (Provine, 1971) of the scientific controversy in Britain that pitted the "biometricians" (led by Karl Pearson) against the "Mendelians" (William Bateson in primis) for more than a decade and which was upsetting British eugenicists, who were fully engaged in convincing British politicians of the merits of their concerns.

One hundred years after its publication, this article and the innovations it contains also occupies a central place in a series of scientific controversies, some of which have important social implications: the utilisation of the measure of heritability, "genes of intelligence", etc. These controversies have also resulted in strong opposition within different scientific fields regarding the level of acceptable simplification in modelling living systems and intergenerational transmissions. In spite of all this, we have lost count of the number of works published in the most important scientific journals that use the polygenic model (or its permutations) without any discussion. The modelling tools proposed by Fisher are used with such frequency and in such a systematic way as to have developed true versatility. Used routinely, they have become invisible in a way and have simultaneously acquired a strong "social robustness": they can be called upon in different contexts, without having to justify their use. Their use is so self-evident that the modelling work and the assumptions that underpin it are no longer truly apparent. Criticism is difficult and often confined to scientific settings whose debates are rarely publicized. The emergence of alternatives is complicated by the hegemonic position of the model. However, the latter imposes strong constraints on the framework of researchers’ thinking and their forms of reasoning, which are not without consequences on the evolution of the disciplines involved and on the social effects of the knowledge produced.

While the production of Big Data and the conceptual issues surrounding their analysis involve more and more researchers in the life sciences and biomedicine, this workshop proposes to take advantage of the centenary of the publication of this article to lead a multidisciplinary reflection associating researchers in the social sciences (philosophy, history and the sociology of science) and researchers in genetics from different fields (evolutionary, animal, plant, human).

Our goal is to return to the text by comparing socio-historic research on the conditions of its creation, on its trajectory and that of the tools it has suggested, with analyses on its present uses - including its “avoidance” when the text is cited without specifically dealing with its content – and the controversies it can provoke. 

The objective of this work is thus to create a space for critical and innovative discussion, informed by the most recent developments in the various contributing fields: evolutionary genetics, human genetics, history and the sociology of the sciences.

[1] Google scholars thus counted 4524 citations since 1978 – the number of annual citations nearly quadrupled between 1991 and 2018.

Copyright © 2020 Cermes3