theories under stress: the evolution of early genetics
Post on 22-Feb-2017
Embed Size (px)
THEORIES UNDER STRESS:THE EVOLUTION OFEARLY GENETICS
The Science of Evolution and the Evolution of theSciences, 13/10/2016
Charles H. Pence
Department of Philosophyand Religious Studies
How can we understand thestructure of scientificcommunities duringtheoretical crises?
THE BIOMETRY /MENDELISMDEBATE
color by Mads Madsden
Fleeming Jenkin, in 1884, from memoir by R.L. Stevenson
William Bateson, in 1905
Karl Pearson and Francis Galton
From Weldon (1893)
St. Cross Church, Holywell
The basic idea: (e.g., Provine 1971)
1867: Jenkins review of the Origin
1892: Batesons Materials
1893: Weldons first biometrical work
1901: Rediscovery of Mendel
1906: Death of Weldon
1930: Beginning of Synthesis
Kyung-Man Kim, 1994
After Fig. 2, Kim 1994
New emphasis: paradigm articulators those whoarticulated the still inchoate paradigms byextending and elaborating the theory, but withoutevaluat[ing] their mentors theory (Kim 1994, 35)
Five of these Darbishire, Schuster, Yule, Pearl, andShull converted from biometry to Mendelismbetween 1903 and 1910.
Kims focus: structures ofeducation, training, and theory
Good! But this is an activedebate in the literature.
Can we detect its signal there?
From previous work (Pence 2011, 2015) Iknew some of this debate played out inNature. Lets find more.
A network of around 100 biologists workingon heredity published around 2,000 articlesin Nature between, roughly, 1870 and 1940.
Aside: Check out the data! The network I will bedescribing can be interacted with live at:
And all data is at:
(data: full network, animated network,time slices)
1894: No robust clustering, standardcenter-periphery network
189599: Cluster of people involved in debatepulled out of broader conversation
190004: Bateson and Weldon completelyseparate from remaining network
190509: Last biometrical analysis,Pearson/Pearl working together; Weldonretreats to experimental work, dies
1910: Back to a cluster-free network
Community structure isreflected in the structure ofthe network of discourse
But! Its not straightforward,and the networks of discoursegive us interesting questionsto ask about the community.
Paradigm debaters?Paradigm warriors?
Participating in debatesbetween paradigms pulls youout of the broader network.
Networks of discourse dontsort paradigm A from paradigmB, nor do they give us Kims
After Fig. 2, Kim 1994, line weight proportional to edge weight in network of discourse, dashed line indicatesconnection present in Kim but missing in new network. Pearson-Pearl line reduced for clarity.
You get a variety of linksacross paradigms, and thoseconnections can be difficult todescribe in any other robust
Problems and Next Steps:
This is just one journal, broadly based in the UK.Cant see Davenports school in the US very well.
Another siloing effect: biometricians found anew journal, Biometrika. Working on dataaccess now.
Just one case study! Need more!