Somebody expressed interest in reading an essay I wrote for my Philopsophy of Science class. Yes, I had a hard time not talking about “singularity” and “artificial intelligence” in it, but I figured it was for the best.
In this essay I will address the topic of anomaly in science, specifically whether or not scientists tend to ignore them as Kuhn says, or seek them as Popper suggests. I will argue that both of them contribute to a reasonable normative view of anomalies and science. However, neither truly accounts for the innate complexity of scientific exploration and the changing nature of it.
First off, a definition of ‘anomaly’. I use it to mean an instance which, at the time, has no peer reviewed, widely accepted scientific explanation. Neither Kuhn nor Popper argue the existence of anomalies in science, or the key role they play – only how they should be approached. Kuhn says in normal science we ignore or do not see anomalies, but to quote his postscript, “If … no one reacted to anomalies or to brand new theories in high-risk ways, there would be few or no revolutions”. Popper believes that Kuhn’s “normal science” is a dangerous idea – while the ignoring of anomalies may occur, it shouldn’t be standard nor accepted practice; scientists should always listen to their data. However, in a TV interview published in 1971 with Brian Magee, his response to “why wasn’t Newton’s theory abandoned with Mercury’s anomalous perihelion in 1820?” was “Because Mercury’s perihelion was not a serious anomaly” (Lakatos 91). This distinction between serious anomalies, and one that can be resolved (or ignored with little consequence) is common to both Kuhn and Popper. The Duhem-Quine thesis argues that the complexity of science makes a critical experiment impossible – so how do you distinguish an anomaly that falsifies a theory, from the kinds of anomalies that don’t?
In hindsight, it might be easy to see a clear cut example of a theory conclusively disproven – for example, spontaneous generation. This was the idea that life is born from inanimate matter – flies born from meat, worms born from dirt. It was a commonly held belief because people routinely saw this happen with their own eyes. Louis Pasteur is generally touted as the man who finally disproved this idea – he even won an award for the design of his experiment. I imagine Popper would correctly identify this is as a good and historic scientific example of falsificationism. To give Kuhn his due, Pasteur did his experiment in 1859 – but the spontaneous generation was first challenged in 1668 by Francesco Redi (Levine and Evers). 200 years is a long time, and lends some weight to Kuhn’s argument that if it were not for scientific revolutions, science would progress quite slowly.
However, I think even in retrospect many anomalies are still difficult to resolve conclusively; for example Lamarck. Lamarck was a French biologist, born 65 years before Charles Darwin. He proposed the idea that one organism can pass on characteristics acquired during it’s lifetime to it’s offspring, as well as that organisms would generally acquire the traits they use, and lose unused traits. A classic example of Lamarckism would be how a heavily muscled animal is likely to have heavily muscled offspring. Under this premise, a German scientist by the name of Weismann proposed that we should be able to see this transmission of the acquired trait of having a short tail. He proceed cut off the tails of many mice and saw that none of the 901 offspring, born from 5 generations of parents with mutilated tails, were born with tails any shorter than a normal mouse (Weismann 432). This presents an anomaly under Lamarck’s paradigm, but even Weismann says his experiment doesn’t falsify Lamarck’s theory. Darwin, even in his proposition of natural selection as the primary driving force of evolution, did not exclude that acquired characteristics might be heritable – and in fact embraced it in The Variation of Animals and Plants under Domestication (footnote: Kuhn talks about the editing of science in textbooks, but despite not being accepted as the primary mechanism of evolution; Lamarck made it into the textbooks, albeit very edited. Ghislen covers this very well.) The shift in general consensus from Lamarckism to Darwinism would clearly classify as one of Kuhn’s paradigm shifts, but that doesn’t answer when, or if, Lamarck’s theories have been disproven. Neither Kuhn nor Popper propose a credible, non-arbitrary way of distinguishing trivial anomalies from paradigm changing ones.
I’ll give one more example that shows even scientists cannot necessarily distinguish between the two. It comes from the area of sleep research, specifically rapid eye movement (REM). Kleitman, considered a pioneer of sleep research, along with his student Aserinsky discovered REM (Aserinksy and Kleitman). However, another student, Dement, was the one who linked REM to dreams (Dement and Kleitman), characterized the different stages of sleep, and eventually started the first sleep lab at Standford. Why didn’t Kleitman pursue REM himself – why isn’t his name the first name on either of these papers? One might say he was making room for the next generation of scientists – but it’s more likely that it didn’t seem interesting or fruitful to him.
Which leads me to my second point – science is increasingly complex. There are an estimated 20 billion neurons in the brain, with an average of 7000 synaptic connections each (Drachman).Somewhere contained in the brain is the mind and consciousness, but how do you approach that? War was one of the major factors in advancing neuroscience – all those soldiers with injured brains coming home, able to articulate the symptoms they were experiencing – much better than monkeys. This progress was made as a result of sociological conditions, and entirely outside the realm of Kuhn or Popper’s outlined vision of science. Let’s put the mind aside though, and pick up something perhaps more manageable – for instance, memory. Back in 1984, picking a theory for how memories are made and how they are retrieved was a matter of creativity and faith (Lynch and Baudry). The only reasonable way to frame anomalies and success criteria was to commit to a theory. The theory is then babied along – there is no possible way to determine whether you’re completely wrong, or your experiment is just insufficient. Questions arise: “How do we visualize the synaptic connections?”, “How can we visualize the synaptic connections as they are being created?” and “How can I keep this slice of rat brain alive long enough to look at it?” (Footnote: I’m describing the lab of Gary Lynch as depicted in “101 Theory Drive” by Terry McDermott). Positive results are truly the uncommon results. In this sense, given your hypothesis is true, the difficulty of obtaining meaningful results is a matter of adjusting things. Many, many adjustments, and the creation of new tools and methods. In this way, the field of neuroscience might work for a decade without a dominant unified theory of memory – and continue to work until a dominant theory emerges.
Scientists have brains, and I would argue that the whole of scientific inquiry could be classified as a complex adaptive system (CAS). When I look down the list of characteristics of a CAS (Heylighen), each one has a corresponding element in scientific inquiry. The scientists are the agents, and each agent has different strengths and weaknesses, and employ the strategies they think will work the best. They keep an eye on what the other scientists are doing to help develop their own strategies. All of these agents are trying to adapt and improve themselves. You might have scientists that migrate into one field from another – bringing in new variation.There is an entire history of science that scientists might learn from, and it affects current scientific strategy. I propose that as science changes, so will the strategies scientists employ – Shapin’s picture of 17th century science is no longer the case. Being focused, methodical, and trying to falsify everything may yield results like Popper says, but if Kuhn’s paradigm changers that reach for the stars never got results, their strategy would have died out.
One counter to science as a complex adaptive system might be Kitcher’s unification theory, and the implication that good science unifies – the more anomalies a theory explains the better. However, there has been an explosion in the number of interdisciplinary fields in science – we used to have biology, and now we have bioinformatics, biotechnology, biogeochemistry, bioelectronics, biomimicry, and so on. I don’t think Kitcher would say we are doing bad science, or doing it wrong. I think Popper would argue specialization is an important part in science as it allows for deeper, more thorough explanations of complex phenomena, and I would tend to agree. However, perhaps this deeper understanding may give way to unification. Even in retrospect it might be hard to tell.
I’d like to make one more note about complexity, and it requires me to revisit Lamarck. Lamarck’s ideas, I would argue, was never conclusively disproven, and actually has recently been shown to have some weight to his idea of acquired inheritance – it just doesn’t look like what he thought it would. It comes in the form of DNA methylation (Bird). Essentially, certain regions of DNA can be turned on or off by the process of methylation, and this methylation can change over the lifespan of an individual and be passed on to offspring. DNA being discovered as the genetic material happened in 1952 by Hershey and Chase. Lamarck had a big idea, without even the concept of DNA to go on, but somehow the science after him continued to develop his ideas. I think this shows the interconnectedness and complexity of science quite well – and it’s unintentional spanning of centuries.
In conclusion, both Kuhn and Popper have contributed to understanding the way in which science happens, but their differences and contentions with the other ones views strike me as a bit like the blind men and the elephant – each touches a different part of the elephant, and each draws a different conclusion. In time, perhaps a more robust picture of science modeled on a complex adaptive system might emerge. Popper’s ideals may have more bearing if we develop the ability to methodically move through every single anomaly ever encountered. Until then, I propose we pick and choose our anomalies to the best of our abilities, however flawed.
Aserinsky, Eugene. & Nathaniel Kleitman. (1953) Regularly occurring periods of eye mobility and concomitant phenomena during sleep. Science, 18, 273-274
Axelrod, R. & Cohen, M. (2000) Harnessing Complexity: Organizational Implications of a Scientific Frontier.
Bird, Adrian. (2002) DNA methylation patterns and epigenetic memory. Genes & Dev. 2002. 16: 6-21
Dement, William. & Nathaniel Kleitman. (1957) The relation of eye movements during sleep to dream activity: An objective method for the study of dreaming. Journal of Experimental Psychology, Vol 53(5), May 1957, 339-346
Drachman, David A. (2005) Do we have brain to spare? Neurology. 64(12):2004-2005
Heylighen, Francis. (1996) Complex Adaptive Systems (Nov 2nd, 2011, 22:10 EMT)
Lakatos, Imre., Paul Feyerabend, and MatteoMotterlini. (1999). For and against method, including lakatos’s lectures on scientific method and the lakatos-feyerabend correspondence. University Of Chicago Press.
Levine, Russell and Chris Evers. The Slow Death of Spontaneous Generation (1668-1859) (Nov 2nd, 2011, 22:00 EMT)
Lynch, Gary. & Michel Baudry. (1984) The biochemistry of memory: a new and specific hypothesis. Science. Vol. 224 no. 4653 pp. 1057-1063
Weismann, August. (1889). Essays Upon Heredity and Kindred Biological Problems. Clarendon Press, Oxford.