Umm, both? I'm both a reductionist and an emergentist. Not only is the whole greater than the sum of its parts, but the parts are greater than the sum of the whole.

Cities arise from buildings, but buildings are made of bricks. To choose only one or the other, reductionism vs. emergentism, would be to limit your perspective either to the small or to the large. I have found that most people who profess proudly, "But the whole is greater than the sum of its parts!" don't really see that that leaves them without a foundation. They propose empty ideas like 'spirit' and whatnot. Without the foundation of a physical substrate, these ideas are impotent. On the other hand, some (few, but some) naive materialists reject the obvious complexities of reality by focusing too much on the word 'just'. We are atoms, yes, but we are not 'just' atoms. We are atoms in a particular arrangement, with particular processes, which is stable through time, which means we are 'more' than 'just' atoms. So we are 'more' than the sum of our parts. But if we ignore the parts, and propose ideas like 'spirit', then actually the study of the parts themselves will reveal more understanding than speculation about bodiless spirits. So the parts are 'more' than the sum of the whole (in other words: the whole, when conceived without relation to its parts, teaches us less than the parts, when conceived without relation to the whole).

I pretty much agree with Hamby.

I might just add some of the other ways of thinking about this 'whole vs. parts' thing.

Groups of Interacting, connected simple components can frequently display properties which cannot sensibly be displayed by the individual bits which make them up. That's pretty much a no-brainer; any function which requires more than one atom cannot be achieved with a single atom.

At a very elementary level, separate piles of sand, cement, and gravel, and a supply of water, cannot be used as is to construct a building, they have to be combined into concrete, which still has essentially the same set of atoms, yet dramatically different physical properties.

You can't access the internet on a pile of unconnected semiconductor chips, circuit boards, wires, etc, even if you have all the same set of bits that a computer is constructed of.

So of course the whole can easily be more, in at least some sense, than the sum of its parts, that is a utterly undeniable. Even if the whole is not clearly 'more', it will almost certainly be an entirely different sort of entity, assuming the 'parts' are actually closely linked to each other in some way beyond simply being in the same disorganised heap.

And of course, as Hamby says, you have to study both the properties of the parts and the way they interact in the composite entity, and how the capabilities and attributes of the 'whole' emerge from the 'parts'. Only a philosopher could generate some pointless argument about the 'whole' being in some way 'just' the sum of its parts. If you want to extend the mathematical idea, maybe something between the 'sum' and the 'product' (as in multiplication) of its parts would be more accurate, depending how they the parts are organized into the 'whole'.

A primary reason why we run into this 'problem' is that even though in principle the ultimate properties of the highest level organised structures are inherent in the simple properties of the ultimate particles and the 'laws' which describe the way they interact, we are totally incapable of analysing the behaviour of more than small numbers of particles without first understanding the systematic behaviour of larger common aggregates, such as atoms, then analysing the behaviour of groups of atoms. When the number of atoms get too large, we have to step back further and think in terms of molecules.

This is simply due to the rapid increase in the possible ways a group of even identical simple entities can interact in systematic and unexpected ways as the numbers involved increase.

And so on, all the way up to living organisms, civilizations, solar systems and galaxies...

Arguing about the 'whole' and 'the sum of its parts' is something only the philosophically inclined waste any time on, AFAICS.

This reminds me of a team building exercise I facilitated a few years back. . . . .  "each individual is part of the whole team"   

 Oh, and Pineapple, add this to your list of times I've been wrong and admitted it.

Thanks so much.

I think that the whole is incapable of being greater* than the sum of its parts, and vice versa. The whole is its parts, and the parts make the whole. They are inseperable. 

*Greater does not seem to be a good term to utilize here. More would be a preffered term, as greater implies that something is better than something else. Being "better" would require subjective input, so the terminology doesn't fly for a question. More might also be subject to this, but it seems less likely to in my eyes.

Re: Protein folding....

*Grabs popcorn and waits patiently for this interesting side topic to resolve itself*

Possibly...

Most definitely.  Information theory is a big part of science, and has a lot to do with biology.  Richard Dawkins is a big fan.

By the way, here's where I got my information.  (It took a couple of hours of digging to remember where I'd read it.)

Monod, Jacques. 1971. Chance and Necessity. New Yor: Knopf.  Vintage Paperback, p. 94

This paragraph is cited in Darwin's Dangerous Idea on p. 196

I wouldn't have considered what natural cited as "additional information" because (a) temperature is constant in the cell and (b) For proteins which require certain pH to function such as those in lysosomes, their ability to shift into their functional conformation when the pH alters is a consequence of their sequence and (c) the ability of proteins to respond allosterically to certain coenzymes and ligands to produce certain functional alterations is again, a consequence of their sequence. These facts are crucial to the functioning of cells. For example, in the central process of protein translocation, for proteins to be imported into membrane bound organelles which are topologically distinct requires them to unfold as the polypeptide chain is passed through, and refold again on the other side. I was thinking more about the heat-shock pathways in which the hsp chaperons detect malformed proteins and refold them, and if they are irreperable, to reject them into the proteolysis pathway. Also, one could argue that the actual conditions specific proteins need to fold are created by other proteins as a consequence of their function which is a consequence of their sequence and their transcription pattern which depends on the presence of GRP which is again a consequence of their sequence. For proteins which must enter membrane bound organelles which are pH specific, those specific pH values are maintained by H+ pumps, which have that function on the basis of their sequence and are present in the cell due to the activity of gene regulatory proteins which is again, due to their sequence. Likewise for allosteric alterations, and the presence of coenzymes altering the function of proteins by mechanically induced structural alternations such as retinal in rhodopsin, their regulation and (except for those that come from the diet) their presence is ultimately a consequence of enzyme pathways and therefore, the activity of those proteins requiring them is determined by other proteins whose activity in turn...um, well you see where I'm going with this.

1971? Jesus, no wonder its wrong. This discipline didn't even exist back then.

This is garbage. Firstly, this is too vague to be meaningful. No serious academic circle would accept a statement involving the phrase "data content of a protein" and they certainly wouldn't accept the phrase "richer" which means nothing.Secondly, the "direct contribution" made to the structure of a protein by the genome is the sequence. Even in 1971, that was known. It was elucidated by Watson in 1958! The sequence determines the structure. The latter statement has been experimentally verified. You can verify it yourself if you want. Take a solution of catalase and add hydrogen peroxide (the substrate) and hook it up to a gas syringe. You'll see gas being given off. That's the oxygen from the decomposition reaction. Take a malforming agent like concentrated sodium hydroxide and add a known amount, which will cause the protein to spontaneously unfold. Now add hydrogen peroxide. No more gas. Now titrate in hydrochloric acid until the endpoint is reached. Add the susbtrate again. The reaction proceeds. Once the malforming agent has been neutralized, the protein spontaneously refolds in vitro. That's it. You've just demonstrated that all the information needed to create the functional 3D structure is contained in the sequence which is contained in the genome.

No it doesn't. In terms of protein encoding (which is a rough gauge of biological complexity, I say "rough" because of things like alternative splicing) the human genome is the most complex, with 30,000 proteins and holds more "information" in the sense that it contains a greater amount of sequence directly responsible for the encoding of amino acid sequences and more complex regulatory regions. It's true that many plant genomes have more base pairs than human genomes but we never talk about Eukaryotic genomes in terms of number of base pairs because it has no correlation with biological complexity. Actually, we never use the term "information". We talk about coding regions, regulatory regions, STR regions, etc. but never "information".

It would still be true to say that a longer genome may have more information content, using the simplest measure of information content, namely the shortest number of bytes which could encode the sequence of base-pairs using the most efficient compression possible. This would depend on both the number of base-pairs and the amount of repetition and redundancy in general, eg a run of 2000 identical base-pairs could be described by no more than 14 bits, ie no more than two ASCII characters. This assumes a simple run-length compression scheme.

If we are describing a genome by the number of active encoding sites, then yes that would be a more meaningful measure, at least according to the assumption that non-coding or 'junk' DNA regions have no utility at all.

This begs the question to an extent, since there are hypotheses, and some indicative studies, that they may serve some use, perhaps as a repository of 'dormant' genes which some mutation or relocation could activate and serve as a potential new step in evolution. Or perhaps some other regulatory function, or structural purpose. So maybe some allowance should be made for these regions. It would be exceedingly hard, perhaps futile, to quantify this in hard measures of 'information content'.

Assessing information content of such real-world structures is somewhat fuzzy, but we can usually put useful upper and lower bounds on it. The length of a lossless compressed bit-sequence which could encode the raw sequence would be set an upper bound. The number of unique functional proteins would define a lower bound.

This is why the subject is so difficult to assess, and in fact why biologists in general shy away from these terms. Yes, we can define DNA as a three-bit base-four system because in terms of digital-to-digital conversion (a sequence of nucleotides to a sequence of amino acids) that's what it is. But because storage of amino acid sequences by DNA is an active process (in the sense that the process of digital conversion is under the control of the very products of the encoding process) it is difficult to assess the biological complexity that could arise solely on the basis of this criteria. Ultimately, whereas the DNA encoding process is a digital to digital conversion process, the actual process of specification of the current state of the cell by how the cell responds to its external signals and internal regulation, both of which are determined by the genome, is an analogue process (that is, possible states are continuous, and are defined by the expression pattern of the genome) the process by which the actual digital encoding process is controlled by the very products of that process is an example of digital-to-analogue conversion. That was exactly what was elucidated by me in the link above. So, instead of discussing DNA purely in terms of the number of base pairs (which is not actually meaningful anyway since the system of three-bit base-four encoding applies only to exons, which make up a tiny fraction of the genome itself, and only sequences which have functioning promoter regions could possibly be transcribed anyway) we should gauge the process by the complexity of the combinatorial control of gene expression and protein regulation which allows for the complexity of eukaryota in the first place. The easiest way to do this is to look at the number of coding sequences because we can then roughly assess the extensiveness of the gene regulatory network. The other way is to look at the complexity of the regulatory motifs that lie in the promotor regions of various genes.

First of all, that's not the definition of 'information' I was using. I would use the compressed size of the base pair sequence. In other words, how many bits would be used to store it on a computer. This is roughly approximate to Kolmogorov's measure of information.

Second, regardless whether measured by base pairs or protein coding regions, it is still true that there are many plants with more protein encoding regions than humans. For example, rice have approx. 51,000 protein coding regions, according to here.

Third, when you get into alternative splicing, you are now talking about proteome complexity, not genome complexity. Small amounts of information (genome) can give rise to much greater amounts of information (proteome) via processing (such as transcription and splicing). However, that does not contradict the fact that the initial source of complexity (genome) has a small amount of information to start with.

The very simple equation z_n+1 = z_n² + c of the Mandelbrot set can generate an infinite amount of information, simply by algorithmically following the progression for various starting c values. However, the equation itself can be represented by fewer than a couple dozen bytes. The equation itself has very little information in it.

Likewise with DNA. It gives rise to great complexity and vast information, but a human genome itself can be stored on a single 700 megabyte CD (approximately). If compressed, it could likely be stored in a few dozen megabytes, since it contains a lot of redundant sequences which are easily compressable using standard compression algorithms, such as Zip.

Frankly, it's irrelevant what terms you use in your field. *I* was talking about information in the genome, not biological complexity, coding regions, or any other thing. You're changing the subject by bringing in these other ideas which do not, in fact, contradict my point.

Just out of curiosity, do you have a link to a source that unambiguously shows that humans are more biologically complex (in terms of intracellular contents) than other eukaryotes? I'm not talking about brain complexity, but cellular complexity.

Well, you have to study BOTH "aspects" of the universe to understand it. Though I do think we can discover a lot by looking more considerably at "small" things.

Wow, I have to say I understand very little of what you guys are talking about, I think I'll just answer the original question for now.

Of course the sum is greater than its parts. I can't think of a time when it wouldn't be. It would be interesting if someone could write a conherent consession on this. The thing is that to truely understand the universe (or anything) we must study the parts, ideally reduced as far as possible, and then look at how the interactions create emergent properties. If we keep doing this we can build our knowledge from the ground level all the way up to, well, I suppose the universe in its entiretly.