Article
Introduction: Life, broadly construed
I would like to begin by justifying why we might wish to take a broad definition of 'life'. The first reason is something of a philosophical old chestnut: the "fine-tuning" problem. The argument goes, roughly, that life as we know it could not exist if any of several physical constants, such as the gravitational constant G or the electron charge e, differed even slightly from their observed values. Thus there is (it is claimed) a need to explain the extraordinary luck that the laws of physics are 'fine-tuned' to support life. Two popular explanations are anthropics and God; I have some sympathy with the former and none at all with the latter. But even the anthropic explanation (usually based on inflationary or multiverse theories positing many 'universes' each with its own values of the various constants) is incomplete, for most versions of it assume that observers will conform to some narrow, parochial definition of 'life' — an example of so-called "carbon chauvinism". Even those who postulate, say, organosilicon or methane-solvent biologies do not go far enough; there are processes which look 'physical' rather than 'biological' but which nonetheless exhibit lifelike characteristics, such as helical structures in plasma which can replicate and evolve.
The second reason is derived from evolutionary theory. It seems unlikely that life on Earth runs the full gamut of possible life, for natural selection produces lifeforms optimised for their environment, and it is unlikely a priori that the range of environments on Earth covers the full gamut of environments in which life can exist. Thus, even if life of a kind suited to non-Earthlike environments were to arise on Earth, it would not survive long.
Finally, the third reason is much less well-specified, and is something of an appeal to the general principle of mediocrity. It's simply parochial and near-sighted to assume that 'life' must necessarily look familiar to us.
So there is reason to consider life as being an extremely varied phenomenon, and define it in some suitably broad terms, such as "self-replicating structures with a genotype which determines their phenotype and is susceptible to mutation and selection". By this definition, for instance, biological viruses are considered as life, albeit parasitic, for the host organism is merely the 'environment' for which they have become adapted by evolution.
Parasitic Life? The Alu sequence
Most genomes of life on Earth consist chiefly of so-called "junk DNA", a catch-all term for sequences that don't code for proteins. Some of these sequences are still of functional value (for instance, markers and controls that direct the various synthases and transcriptases), but others are essentially 'noise'. One kind of nonfunctional DNA is the transposon, or transposable element. Transposons (including the subclass of retrotransposons) account for almost half of the human genome. A transposon such as the Alu sequence suborns the copying mechanisms of DNA and RNA to copy itself, not only into the same place in the genome of the next generation (which any intron can do) but also into other locations in that genome. It is a kind of Selfish DNA, which does not (directly) contribute to the reproductive success of its host organism, instead concentrating on its own reproductive success within the host organism's genome.
Transposons are capable of undergoing evolution, for both of their replication procedures (normal genome replication and copy-and-paste repetition within the genome) are susceptible to mutation, and selection pressure is provided by the host's attempts to prevent them (such as RNA interference) as well as by the need to avoid having too detrimental an effect on the host organism's fitness. The latter probably explains why we see no genomes containing a transposon so viral that it overwrites large parts of the host genome with copies of itself — for such a transposon would lead to the host producing offspring with so much of the genome deteriorated that they would be probably stillborn and almost certainly unviable. The transposon population within a genome can be seen as 'competing' for a limited number of sites; they can only safely overwrite other junk DNA, and too much insertion leads to a larger genome which imposes fitness costs on the host. Some of these sites are more 'fertile' than others, for the efficiency and efficacy of transposon replication is affected by the DNA on either side of it. Indeed there are many niches within this environment to which different and diverse transposons are suited. It is evolution writ small, and it runs with even more massive parallelism than does the evolution of species. It is a complicated and complex business, for a transposon has many layered modes of reproduction (the reproduction of the host, copying within the host's genome, horizontal gene transfer, and perhaps others) all of which exert their own selection pressures, and all of which involve the possibility of mutation. One can even consider transposons to be cousins (taxonomically if not genealogically, though the latter is also a possibility) of viruses, plasmids, and the like. Indeed, they are grouped together as the "mobilome"; in eukaryotes the main mobile genetic elements are transposons, while prophages and plasmids are more important in prokaryotes.
Under a broad definition of 'life', we have no choice but to include selfish DNA as a (parasitic) lifeform, separate not merely in species but in substrate from its host organism. The Alu sequence is alive.
Digression: Origins of Transposons
It is fascinating to speculate on the possibility, alluded to above, of genealogical relationship between different kinds of mobile genetic elements. While some transposons could be created by chance mutation (it is hypothesised, for instance, that Alu was formed by the fusion of two similar pieces of junk DNA, originally mutated copies of the 7SL RNA gene), it seems at least possible that a retrovirus, by infecting a part of an organism used for reproduction (eg. the nucleus of a single-celled lifeform, or an egg or sperm cell of a sexual multicellular organism) could be turned into a retrotransposon; at least some of the machinery of replication is, or at least could be, shared by both. Of course there is a similar possibility in the other direction.
But here I must leave such speculations to those trained in the relevant fields of biology, and return to my main topic.
Ethical implications: Genetic Rights?
If we consider all life to have ethical value (not that we necessarily must do so), and if we classify transposons and other mobile genetic elements as life, then it is a simple syllogism to conclude that transposons have ethical value. Of course our values are gradated: we typically assign higher value to sapient animals than to the 'merely' sentient, lower to plants, and lower still to viruses. We don't tend to pronounce the immorality of subjecting plants to harsh conditions or invasive experimentation, nor do we have qualms about causing the extinction of a species of bacterium or virus that is responsible for a disease in humans. But at the same time, we frown on wanton maltreatment of plants: even though they (to the best of our knowledge) do not feel pain, we don't go around snapping branches off trees just for fun; nice, well-behaved people just don't do that kind of thing. Though of course, if we are cold we have no problem with killing a tree and burning its corpse for warmth.
I don't mean to argue that morality depends on the definitions of English words. It's not strictly the definition of the English word 'life' that I care about, but rather the exploration of my utility function, and whether my preferences are consistent and coherent, or whether they make an arbitrary distinction between "life with moral status" (people, chimps, and kittens) and "life without moral status" (cockroaches, E. coli, and transposons). [This paragraph was added in response to Nisan on LessWrong]
We have, then, a complicated (some might say arbitrary) moral code for trading off the values we assign to different kinds of life. But what ethical value do we assign to a transposon? This is not an entirely irrelevant question, since genetic engineering may give us the power to edit transposons out of genomes, and there may be benefits to doing so.
Good examples of this are given by wheat and maize, food crops whose genomes contain large proportions of retrotransposons. It is conceivable that removing the junk DNA from these crops might indirectly improve yields, or have some other beneficial effect such as increasing disease resistance. When our understanding of genotype-phenotype linkages is better, we might be able to modify existing transposons to produce desired mutations, but doing so might involve destroying the transposon's ability to replicate itself within the genome, essentially sterilising it (though it would still be replicated with the genome, like any intron).
In humans, too, there are reasons why we might wish to rid ourselves of Alu and other transposons (supposing, for the moment, that the genetic modification of humans is considered acceptable). Alu insertions can cause the mutations responsible for heritable disorders such as haemophilia; Alu has also been linked to cancer. It is unsurprising that transposons can cause such damage; if they interfere with the genome copying process in cell division, they could easily lead to disruptive mutations.
It is not my intention here to positively argue for the removal of transposons from genomes, whether of maize or men. Rather, I sought to justify the assertion that it is in principle possible that we might have reasons to do so, in order to raise the question of how we would trade off our values in other areas against the intrinsic value of life when that life takes the form of a transposon.
I think that, purely as a practical matter, we cannot concern ourselves with the destruction of individual copies of a transposon. For, cutting down and burning a tree kills one tree, but it kills trillions and trillions of copies of every transposon in that tree's genome. The same goes for killing and eating a cow, or harvesting, brewing and drinking grain. Even picking at a scab on our bodies ends the lineage of more copies of a transposon than we can count. And of course our immune system, when it fights invading diseases, is destroying mobilomes in their millions. No, we cannot worry about the value of individual copies, or else we should have to shut up and multiply, and reach a truly absurd answer.
There is, independently from the practical argument, a more abstract reason not to attach value to the individual copies of a transposon, at least within my philosophical framework, Syntacticism. This is because Syntacticism ignores instantiation. An example of the kind of thing I mean, is that if we have two identical AIs inhabiting identical virtual worlds, switching one of them off has not killed anything. However, if in the future they would have received different inputs, thus causing them to diverge, then switching one of them off has killed that part of them which would be embodied in that divergence. (Actually, a full treatment of this proposition is extremely complex and confusing, involving some heavy-duty possible-worlds reasoning)
The excising of a transposon from the genome of an entire species could, perhaps, be considered roughly equivalent to killing a single macroorganism, though probably painless plant rather than sentient animal. Excising a transposon from the genomes of all life on Earth would be tantamount to causing the extinction of a species — and we were happy to do that to smallpox, but we wring our hands over the dodo. Presumably, if we could costlessly eradicate a harmless transposon, we would choose not to, as that would be pointless destruction — whereas an even slightly harmful transposon could expect no mercy. Then again, a harmless transposon could conceivably mutate into a harmful one; would this risk be considered sufficient grounds for eradication?
Conclusion
Understanding how our ethics cope with the confounding and strange in the form of transposons is not only an interesting question in its own right, it also prepares and practices for the difficulties we are likely to face should we ever encounter alien life, for the odds are that it will not be biological, and that its nature will be at least as strange and jarring as the bizarre world of transposons. If a swarm of sapient plasma helices were to come visiting, how would we respond? Would we behave like ramen, or varelse? Alu, and the other denizens of the mobilome, help throw the relevant facets of our morality into sharp relief.
