The Utility of Design Analogies

Researchers at the University of Edinburgh have discovered a signaling circuit of highly flexible nanotubes in eukaryote cells. From the news release:

Professor Mark Evans, of the University of Edinburgh’s Centre for Discovery Brain Sciences, said: “We found that cell function is coordinated by a network of nanotubes, similar to the carbon nanotubes you find in a computer microprocessor.

“The most striking thing is that this circuit is highly flexible, as this cell-wide web can rapidly reconfigure to deliver different outputs in a manner determined by the information received by and relayed from the nucleus. This is something no man-made microprocessors or circuit boards are yet capable of achieving.”

Scientific progress has not been kind to David Hume’s criticism of the analogy argument for design. At the anatomical level, which Hume observed, any similarity between life and human technology can be hard to discern. But as science progressed and the electron microscope has allowed us to peer into the cell, the analogy has returned with a vengeance. Where would our modern understanding of cell biology be without concepts like regulation, control, signals, receptors, messengers, codes, transcription, translation, editing, proofreading, etc.?

You might even say that science has been well-served by a healthy dose of methodological designism.

This news release was discussed on the Peaceful Science forum, where someone posted a link to philosophers Massimo Pigliucci and Maarten Boudry’s article, “Why Machine-Information Metaphors are Bad for Science and Science Education”. From the abstract:

Genes are often described by biologists using metaphors derived from computational science: they are thought of as carriers of information, as being the equivalent of “blueprints” for the construction of organisms. Likewise, cells are often characterized as “factories” and organisms themselves become analogous to machines. […] In this article we connect Hume’s original criticism of the living organism = machine analogy with the modern ID movement, and illustrate how the use of misleading and outdated metaphors in science can play into the hands of pseudoscientists. Thus, we argue that dropping the blueprint and similar metaphors will improve both the science of biology and its understanding by the general public.

First of all, it is interesting to note the motivation behind wanting to abolish “machine-information metaphors” from science: It “play[s] into the hands of pseudoscientists” – i.e. people who are friendly towards intelligent design. In other words, this is a political call-to-action to stop giving aid and comfort to the enemy.

Second of all, the authors choose to criticize the “blueprint analogy”, which is indeed not a very helpful analogy. The human genome contains nothing that can be likened to a drawing of a human. It contains the instructions for synthesizing the proteins and ribozymes which are necessary for the development and functioning of the human body, but likening this to a blueprint isn’t a very helpful analogy.

The “genome as a blueprint” is an example of a faulty analogy, comparing an analog object (a blueprint) with what’s really a digital technology (encoding information as a symbolic language). As our technology has improved, closer approaching the sophistication found in cells, our ability to formulate accurate analogies has improved as well.

Ironically the blueprint analogy fails, not because design analogies in general fail, but because it tries to establish an analogy to a technology that is insufficiently sophisticated to measure up to that found in the cell.

Unlike the blueprint analogy, analogies to computer technology continue to generate heuristic dividends for science. Does anyone believe science understanding would be improved if we had to describe, say, protein synthesis without using concepts like code, sequence, transcription, translation, etc.?

Using Design to Make Predictions, Redux

The claim is often made that teleological explanations are inherently incapable of leading to predictions, as whatever is observed can be explained with reference to the designer “working in inscrutable ways”.

I’m not terribly impressed by this argument, for one reason: I have successfully used teleological reasoning to derive predictions from.

Years ago, I started looking at homeobox genes from a teleological perspective. Homeobox genes play a key role in development, initiating the developmental cascades that lead to structures such as eyes, hearts, and legs. Homeobox genes have been referred to as the “genetic tool kit” and shows suprising homology across Metazoa. For example, the homeobox gene distal-less regulates both the legs a fly, the paws of a mouse, and the tube “feet” of a sea urchin, even though all of these structures are thought to have evolved independently of each other.

Similarities of homebox expressions in banan flies and mice. Creative Commons Zero, Public Domain Dedication via Wikimedia Commons.

In 2005, homeobox genes were thought to have been present in the last common ancestor of Bilateria (animals which are bilaterally symmetrical). But using the teleological notion of front-loaded evolution, I hypothesized that the genetic tool kit of homeobox genes were even older than that, predating the origin of multicellularity. As I wrote, in August 2005:

If we assume that eukaryotes were designed with the purpose of giving rise to multicellular organisms, we can make certain predictions. For one, we would expect the first eukaryotes to have contained a predecessor to the modern tool kit, and it’s possible that some unicellular eukaryotes still possess it. It will probably not be the full set possessed by modern organisms (or rather, full sets, as several organisms differ in the number of genes they have), as some genes may have been generated through gene duplications, but I definately [sic] expect genes that are clear precursors to modern tool kit genes to be found in unicellular eukaryotes.

I have not been keeping up with the literature for the last several years. However, in checking up on things, I have made the pleasant discovery that my prediction has been fulfilled.

In a paper titled “Homeodomain proteins belong to the ancestral molecular toolkit of Eukaryotes”, the French researchers Romain Derelle, Philippe Lopez, Hervé Le Guyader, & Michaël Manuel present evidence that the last common ancestor of Eukaryotes, the “Ur-eukaryote”, contained genes for homedomain proteins.

The researchers found that homeodomain proteins “are present in all eukaryotic lineages containing multicellular organisms, and absent in exclusively unicellular lineages.” In other words, homeodomain proteins do not appear to play an important role in unicellular organisms.

But the researchers carried out a phylogentic analysis and concluded that the multicellular organisms with homeodomain proteins (animals, plants, algae, and fungi) had all inherited them from the uicellular “Ur-eukaryota” while those lineages who remained unicellular lost their copies through reductive evolution. The findings make the researchers suggest that “eukaryotes as a whole are preadapted for multicellularity”:

As a corollary of ancestral molecular complexity, Ur-eukaryota probably possessed many of the good building blocks, which were subsequently recruited, by convergence in several lineages, to perform the functions required for development of multicellular organisms. In other terms, we suggest that the eukaryotes as a whole are preadapted for multicellularity, which only means that the ancestral complexity of the eukaryote genome and cell biology facilitated multiple acquisitions of multicellularity.

As these findings were published two years after I made my prediction, I feel pretty good about the heuristic value of a teleological perspective.

References

Derelle R., et al., 2007, “Homeodomain proteins belong to the ancestral molecular toolkit of Eukaryotes”, Evolution & Development, 9(3):212-219