Essay: Symbiosis/Cooperation in Evolution/Life
Robert Maas
but it didn't appear, trying again now...)
There is no such thing as 'life' per se. There are various processes
which occur, which we lump together as 'life' processes. The ensemble
of such processes we call 'life' but it's not an object, the word is
somewhat a misnomer, leading to false word-arguments sometimes.
The most fundamental symbiosis/cooperation, where one helps another, is
between identical copies of genetic material. In the important sense,
what is helped is the pattern, which is the same among all the
identical copies, not one or another individual copy/instance of that
pattern. It makes no sense to say that a particular copy/instance of a
pattern is helped when some creature replicates, because replication
creates OTHER copies, OTHER instances of the pattern, never again any
particular instance again. Only the mathematical pattern (incarnated in
matter) is replicated, not any particular instance or subset of
instances of that pattern. Accordingly anything which helps particular
genetic material is helping all instances of a pattern equally, thus
may be most appropriately described as helping the pattern itself
rather than any specific instances of it. Any two different instances
(identical copies) of a pattern are irrevokably interlocked in a
symbiosis whereby anything which helps one (to survive) is equally
helping the other insofar as it is the common pattern which is really
helped to survive. Any attempt by one instance to survive longer is
equally helping all other instances. Any help that one instances
provides to another is equally well helping itself in that it helps the
pattern itself to survive. Accordingly, anything which tries to survive
is cooperating with all identical copies of itself in the sense of
aiding the survival of those other copies because it's really aiding
the survival of their common pattern.
We might add the phrase "in nature" or "in the Universe" to most of
what we say, such as "the fundamental symbiosis in nature", but that's
as redundant as saying "Little Bo Peep's sheep will come back wagging
their tails behind them" (what other tails could they possibly wag
other than their own tails, and where else do you think they would be
wagging them, in front??), because as far as we know "nature" refers to
the whole Universe, and there is no other place that we know anything
about except the Universe.
Symbiosis between identical copies is absolutely 100%. A lesser (<100%)
but likewise irrevokable symbiosis is between different genetic
material which gets converted back and forth by random mutations faster
than the characteristic time of evolution. Thus we may correctly speak
loosely not just of the survival of one specific pattern but of a
cluster of similar patterns which get converted one to another rapidly
enough that over sufficient time none of the individual patterns are
conserved but the cluster of nearly identical patterns is preserved.
The entire cluster of interconvertible genetic material may correctly
be called a 'gene' rather than calling one or another of the specific
patterns a 'gene'. Note that this applies ONLY to patterns that have
reproductive advantage, not to patterns that are involved in 'neutral
mutations'. The center of such a cluster would be those patterns which
convey the particular kind of advantage, while the periphery would be
those other patterns which are similar enough to be created from the
central patterns, then by chance survive (despite loss of reproductive
advantage) long enough to be converted back to a central pattern before
they die out. Patterns too far from the center of a cluster would die
out before they would by chance mutate back to the center, and are thus
properly considered distinct mutations of the gene (cluster) rather
than members of the gene (cluster).
Of course this definition of which specific patterns are members of the
cluster, and which aren't, is fuzzy. We can't say precisely whether a
particular pattern on the borderline is or is not a member of the
cluster, i.e. whether it is to be considered an example or not an
example of a particular gene (cluster). Likewise if the periphery of
two different clusters overlap, it makes no sense to try to pin down
whether a particular pattern is to be considered a member of this gene
(cluster) or this other gene (cluster). The best we can say is that the
specific pattern is a member of BOTH clusters, perhaps tighter in one
cluster than the other because of a higher rate of mutations to/from
the center of this cluster than that.
I define 'gene' as the pattern of any segment of reproductive material
that is long enough that random mutations don't spontaneously create it
repeatedly during evolutionary time, so that if it's lost it in general
cannot be replaced, yet short enough that it usually takes longer for
recombination or mutation to change it than the characteristic time of
selection for its phenotype. Thus a sequence of ten base pairs is too
short to be considered a whole gene, while an entire chromosome or DNA
loop is too long to be considered a single gene.
Although there are specific boundaries between codons, that is between
the set of bases which code for one protein and the nearly (*) adjacent
set of bases which code for a completely different protein, there are
no specific boundaries between genes, that is between evolutionarily
conserved and hence evolving pieces of genetic material. If any
particular length of genetic material (specific pattern sequence or
cluster of such) contains sufficient reproductive advantage that it can
maintain effective fecundity of one or more over a long time, chances
are that a slightly longer piece including that one, or a slightly
shorter piece included in it, would likewise maintain fecundity of one
or more, perhaps over a slightly shorter time but nevertheless long
enough to be correctly considered 'evolutionary time'.
(*) Because of non-coding regions between adjacent codons, two
'adjacent' codons may not be precisely adjacent segments of DNA. But if
the non-coding regions are sufficiently small compared to the two
codons, the frequency of crossovers (assuming sexual reproduction)
which occur in the non-coding regions and thus disassociate the
'adjacent' codons, would be less than the frequency of crossovers
interior to one codon or another which effectively mutate that
particular codon. So if a cluster of variations of each codon has
sufficiently large evolutionary advantage as to achieve raw fecundancy
sufficient to overcome all these interior crossovers and thus achieve
effective fecundancy larger than one, the combined raw fecundancy of
the two codons together is almost surely enough that the combination of
the two codons achives effective fecundancy greater than one (i.e. the
two codons stay together long enough that the combination of the two
together continues over evolutionary time, thus the combination may be
considered a single gene as correctly as considering each single codon
a gene itself).
Accordingly we may correctly consider a single codon to be a gene, or
two or more adjacent codons to be a single gene, or just part of a
codon to be a gene when that part of the generated protein itself
conveys selective advantage somewhat independent of other parts of the
same protein, and finally we may consider just about any length of
genetic material overlapping several codons and some non-coding
sequences and parts of a codon at each end to be a single gene,
whenever that particular segment of genetic material happens to have
effective fecundity greater than one (as a pattern, or as a cluster of
similar patterns). In a segment of genetic material that conveys great
selective advantage, a wide range of lengths may simultaneously be
considered to be genes. Lengths too short, because by pure chance
identical sequences exist in other unrelated places, such that the
fecundity achived here is not sufficient to overcome random mutation of
OTHER copies away from this pattern or cluster of patterns (remember
it's all copies of the pattern, not just these particular copies in
this place which must be counted before and after when computing
fecundity), would have effectively fecundity less than one (they
diffuse i.e. mutate away faster than they are restored any time they
happen to start with greater than chance frequency), thus not properly
be considered a 'gene'. Lengths too long, because they get broken up by
crossovers faster than their combined raw fecundity can make copies of
them intact, are also not properly considered 'genes'. But any length
between these two extremes, where selective advantage overcomes these
two effects, could correctly be called a 'gene'. Thus, while it's
appropriate to consider any such length of genetic material to be a
'gene', it's not appropriate to define the start and end of a gene and
argue over whether this or that piece of genetic material is one gene
or two genes or part of a gene. Rather it's appropriate to allow the
word 'gene' to apply to ANY length of genetic material which conveys
sufficient selective advantage that its effective fecundity (in some
environment, not necessarily the one under discussion) is greater than
one, and to allow for cases where we simultaneously speak of one gene
overlapping another or one gene being part of another.
Note that there's a fundamental symbiosis between any two genes that
overlap, in the sense that many of the aspects of selective advantage
that cause one to have fecundity greater than one hence be termed a
'gene' would happen to be caused by some overlapping codon(s) or
sub-codons and hence be shared between the two overlapping genes. I.e.
if a particular gene is, by means of the overlapping potion, enhancing
its own survival, it is simultaneously enhancing the survival of the
'other' gene which overlaps it.
So-far I have discussed symbiosis as evolutionary cooperation between
identical copies of genetic pattern, between similar patterns which are
interconverted fast enough that they survive together rather than
individually, and between overlapping patterns/clusters. The result
so-far is a hierarchy based on containment (in a cluster, or in a
longer piece of genetic material), whereby small pieces of genetic
material (small genes) are surviving partly because of their own
efforts and partly because they are part of larger evolving entities
(larger genes). It's not proper to say that evolution occurs only at
the small level or only at the large level, or that evolution occurs at
both levels but one level of evolution is merely a consequence of
evolution at the other level. Rather it's proper to treat several
levels (sizes) as evolving simultaneously, in fundamental symbiosis
with each other. Evolution occurs at the individual codon level, and at
the cluster of similar codons level, and at the larger segment of
genetic material level, all simultaneously and in symbiosis with each
other. This symbiosis is based on two genes being nearby on genetic
strands, or even overlapping (using my loose definition of 'gene' which
permits overlapping segments to each be called a 'gene'), and of course
also based on any two identical or sufficiently similar (same-cluster)
sequences regardless of their physical distance.
The generalization of this symbiosis to different genes enclosed within
the same cell, then to different cells included within the same
organism or colony, then to different free-living cells or organisms
which share a local environment, is obvious. Evolution is a symbiosis
between patterns and clusters of patterns of genetic material 'trying'
to out-survive the alternative patterns & clusters, cell lines 'trying'
or actually trying with intent to out-survive other cell lines except
those which share sufficient genetic material that "selfish gene" or
"prisoner's dilemma" (*) cooperation overwelms cellular competition,
and societies of organisms 'trying' or deliberately trying to
out-survive other societies except where cooperation is more beneficial
than competition.
(*) When two genes are sufficiently interlocked in cause/effect
regarding survival, such as being forced to live together for a long
time within cells or habitats, a game-theoretic analysis often shows
that it pays more to cooperate with each other for a while (so long as
the interlocking is maintained) than to be directly and harshly
competing with each other. In borderline cases where neither
unconditional cooperation nor unconditional warfare is appropriate, a
"tit for tat" strategy, such as in the "prisoner's dilemma" game, is
often optimal.
Evolution of course also applies to emergent qualities which Richard
Dawkins calls "memes", intellectual concepts which aid the survival of
anyone who has those concepts and effectively uses them. Because memes
are emergent (they can't exist without some substrate life-form), memes
are totally dependent on the underlying life-form, but because memes
enhance survival of the life-form, life-forms are somewhat dependent on
the memes. Evolution is thus a symbiosis between the genes and memes
together. A conscious living being would thus, if wise, choose to work
for the survival of all his/her genes (and copies of same in other
living beings) and simultaneously all his/her beliefs (memes), except
might choose to abandon some genes and/or memes which interfere too
much with the survival of the rest, but being careful to allow for
error of judgement by purging a gene or meme ONLY when it is
conclusively rated as grossly harmful, and even then to perhaps retain
an inactive copy of the purged gene or meme as backup in case the
purging turns out to be a mistake. For example, genocide and
book-burning are bad, but differential survival where genes die out
only over a long term is fine. Even the Smallpox virus, extinct in the
wild now, is being preserved as a list of bases in a computer database,
and there's debate whether the last archive of the physical virus
should be retained or destroyed, because we might someday need it again
for some research or medical purpose.
Although cooperation on a global (worldwide) scale is essential to the
survival of any particular life on Earth, evolution does NOT apply to
the world as a whole because at present there is no alternative. (With
due respect for and appreciation of the other fine work by Lynn
Margulis and Stephen J. Gould, their concept of 'Gaia' is nonsense, and
they are diminishing their credibility by incessantly advocating that
nonsense. The Earth is in essense NOT the same as a living organism.)
But if and when we expand life far beyond Earth, where we have more
than one planet-size ecology, sufficiently independent that they can
effectively compete with each other for survival in this VERY BIG
GALAXY, and perhaps beyond in other galaxies as well in this INCREDIBLY
BIG UNIVERSE, then it would be appropriate to extend the concept of
evolution by natural selection to planets or solar systems as a whole,
and to consider each individual planet or solar system to be a living
entity, essential each a collosal society competing with societies on
other planets or in other solar systems. Then symbiosis/cooperation in
life/evolution would include symbiosis between planets/solarsystems and
their inhabitants and cooperation between different
planets/solarsystems which shared parentage or other combinations of
genes and memes.
(Before finding time to post this essay, I've been pondering a related
but distinct topic, and I've now (1999.Jan.17) added this topic to the
essay:)
A rather different question is the basis of DIFFERENTIAL survival. For
genetic material, it seems the SNP (Single Nucleotide Polymorphism),
inm the context of the gene it is within, is the answer in most cases.
A chance mutation creates a different nucleotide at one place within a
gene. There are now two different genes completing with each other.
(Allowing overlapping genes, we have two families of genes, those with
the particular mutation and those with the original nucleotide at that
position.) If the mutation is nearly neutral, we have a cluster of
similar sequences that constitutes a single gene-cluster rather than
separate competing genes. But if the mutation is non-neutral,
competition (differential survival) between the two sets of genes will
increase the number of copies of one of the genes at the expense of the
other (possibly with different results in different habitats, that is
the mutation becomes common in one habitat while the original remains
common in the other habitat), which has the effect of eliminating the
SNP one way or the other (by fixing (*) the original or the mutation),
or nearly eliminating it one way in one habitat and the other way in
the other habitat).
(*) Usual genetics definition: One allele gets 'fixed' if all copies of
the alternative allele are eliminated from the gene pool.
Because of my fuzzy defintion of how long a segment of reproductive
material (usually DNA) constitutes a gene, when two mutations occur
close together, so that for a while there are two SNPs close together,
they may be thought of as lying within a single gene (and thus
cooperating) or lying in disjoint genes (and thus competing) over
different time spans. This is not a problem, in fact it helps to
clarify thoughts about what really happens. If each SNP creates
differential survival, either globally or per habitat, so that neither
SNP effectively generates a gene cluster, then initially (assuming the
two mutations occurred in different individuals from the original,
rather than successively in a single hereditory line) there are three
different alleles of any gene that includes both SNPs: The original,
the first mutation only, and the second mutation only. Until such time
as recombination occurs between the two SNPs within individuals who
happened to be pure-first-mutation and pure-second-mutation, these
three alleles are the only ones competing with each other, and the two
SNPs are locked in a fundamental symbiosis. If one or the other is
disadvantageous, we can expect that single-mutation strain to die out,
reducing the problem to a single mutation. But if both are
advantageous, so that the original gene is eliminated, the two single
mutations will be directly competing with each other until one
eliminates the other or recombination between them creates a
double-mutation gene, yielding a three-way contest again. Eventually
one of the four possible genes becomes fixed globally, or near-fixed
per habitat.
But we can also think of smaller genes where each SNP occurs in a
different gene, then our langauge has us saying that there are two
alleles of one gene and two alleles of one gene, and the two genes are
"linked" because of infrequent crossovers between them, so a symbiosis
occurs between the two linked genes over timespans long enough for
differential survival to be important but too short for recombination
to be important.
Over short timespans, speaking of a single gene with three alleles, or
speaking of two genes with two alleles each and linkage between the two
genes, are equivalently appropriate. Over longer timespans, speaking of
two separate genes that are only temporarily linked until recombination
breaks that linkage, is most appropriate.