If we can use phylomemetics to do the samething for memes,
then we will be on ourway tohaving an actual science of memes.
Tothat end, Ithoughtthat creationistlegislation wouldbea
perfect test case.
One key requirement for phylomemetics is that memes are
large and relativelystable. You need to have dozens or hundreds
of identiiable characteristics to build a good phylogeny. This
is the reason that a phylogeny of LOLcat memes on the internet
would probably be diicult. There are only so many charac-
teristics you can get out of the viral meme “I CAN HAZ
CHEESEBURGERS”.But with legislative proposals we have the complete text,
along with the date that the proposal was published. And,
unlike fossil species, we think we have all of the bills proposed
between 2004–15 in a database of anti-science legislation main-
tained by the National Center for Science Education (NCSE),
which is devoted to defending science education from polit-
ical and religious attacks.
To do the phylomemetic analysis, I took 67 policies from the
NCSE database and spent several weeks lining up the texts,
coding the variants (418 kinds of variation for each of 67 poli-
cies) and then running statistical analyses on the computer.
The main program I used was BEAST (Bayesian Evolutionary
Analysis Sampling Trees), which is the workhorse program
behind thousands of published papers. BEAST specialises in
analyses that infer the timing of events in an evolutionary tree.
If you have ever heard of a evolutionary study that makes conclu-
sions about how many million years ago an event occurred, it
was probably a study that used BEAST.
Biologists use BEAST to study everything under the sun,
from languages to dinosaurs, but BEAST is also used to study
the evolution of diseases like HIV and Ebola. In fact, in 2014,
while the Ebola outbreak was still going on in West Africa, a
paper published inScienceused dated samples of the Ebola virus
to track its spread and estimate when Ebola was introduced
into each new country. These estimates were measured with a
precision of weeks rather than millions of years. A shocking
feature of that paper was the postscript, which noted that ive
of the coauthors of the study – health workers in West Africa
that had helped collect the virus samples – had died of Ebola
before the paper was published.
This goes to show that phylogenetics can be life-and-death
stuff, and this is part of the reason I get annoyed with creationist
attempts to get politicians to pass laws suggesting that evolu-
tion is controversial science that might just be make-believe.
A great feature of BEAST is its speed. In just a day or two I
could sample tens of millions of possible dated phylogenetic
histories of anti-evolution bills. It turns out that these bills
show a strong signal of successive copying – Darwin's “descent
with modiication” – so the summary tree is likely a close esti-
mate of the true history (which we never know for sure).
What did this tree show? The strong signal indicates that a
creationist legislator tended to write a bill by copying from one
previous bill, rather than combining text from multiple bills
or re-writing bills from scratch.APRIL 2016|| 15The Science Education Acts target
not only evolution and origin-of-life
studies but also human cloning and
global warming.enterlinedesign/adobe