Republished from Ontic Cafe 2014
This post is the first of two. The material is from a quite complex field of the philosophy of biology and information theory in biology. Nonacademic philosophers should be able to get a reasonable idea what is going on and benefit from an introduction to one of the hottest topics in the philosophy of information and biology.
Different Conceptions of Information
Let’s start with a rapid introduction to the philosophy of information. There are several conceptions of the nature of information – of what information actually is. These conceptions vary dramatically in their details and ontological commitments – the things that are taken to be necessary to have for there to exist some information (in philosophical language we day “the necessary conditions” for information to exist.) Here a couple of quick examples will be instructive.
The most common understanding, and the most common scientific one, is that of quantitative information theories. In these theories one has information on a statistical or probabilistic basis. According to these conceptions information exists when there is a reduction in uncertainty about what is happening at an information source. An information source is any physical process that can be modeled statistically – about which you can say there is a certain probability of the next state of the source based on the current one. A simple example is you reading this sentence. Each word makes the next word more or less likely because of the structure of the English language and the rules (grammar and meaning) for making English sentences. The source is the text you are reading. This is the very example most used by the founder of modern quantitative information theory – mathematician Claude E. Shannon (The Mathematical Theory of Communication, 1948.)
The main alternatives to quantitative statistical theories are algorithmic theories. These involve measuring the complexity of strings of symbols or what are called data objects. Any sequence of elements can be a data object. The longer and more complex the data object – the more information it has. The most famous is that developed by Russian materialist mathematician Andre Kolmogorov. In Kolmogorov’s theory the amount of information in a data object is given by the length of the program or description required to generate or construct the data object.
Quantitative statistical measure based conceptions and definitions of information have often been seen as inadequate because as Claude Shannon himself wrote in The Mathematical Theory of Communication, they do not attempt to capture any meaning of the symbols that are transmitted. His predecessor R.V.L Hartley wrote that “[i]t is desireable therefore to eliminate the psychological factors involved and to establish a measure of information in terms of purely physical quantities” (Transmission of Information, 1928, 536.)
Shannon’s peer and mentor Warren Weaver first observed that in future it would be desirable to formulate a conception of information that accounted for meaning. Later theorists came to refer to such conceptions as theories of semantic information. There have been several of these – mostly naturalistic – offered by both mathematicians and philosophers. The first notable attempt was by the famous Vienna circle mathematician and philosopher Rudolph Carnap. Carnap joined with mathematician Yehoshua Bar-Hillel to formulate a theory of semantic information in which the semantic information content of a sentence was determined according a to a logical formulation (1953.) In lay terms the information content of a sentence is the set of all sentences that are false if that sentence is true.
Later various other conceptions of semantic information. Philosopher Fred Dretske adapted elements of Shannon’s theory (1981 – Knowledge and the Flow of Information.) Mathematician Keith Devlin produced another logical conception (1995- Logic of Information.) More recently, Luciano Floridi has produced a theory of semantic information that extends and adapts ideas put forward by Devlin and Bar-Hillel and Carnap. It is different in that it requires information to have alethic value – to be based upon data which are truthful according to certain fairly complex criteria (Floridi, Information in The Blackwell Guide to the Philosophy of Computing and Information – 2004, Information – A Very Short Introduction – 2011, The Philosophy of Information – 2012.)
The idea of semantic theories of information is that information and meaning are directly related somehow. Usually meaning is thought to involve truth value of some kind.
Meanwhile in Physics and Biology
An enormous part of the story of our understanding of the nature of information comes from physics. I will not say much about that here, except to say that physicists often regard information to be a physical thing. Another pioneer of information theory – the father of Cybernetics Norbert Weiner – once said that “information is information, not matter or energy…no materialism that does not admit this can survive…” (1962, Cybernetics: or Control and Communication in the Animal and the Machine.) No physicist has claimed that information is matter or energy, but quantum computing pioneer Rolf Landauer was sure that it is physical (Information is a Physical Entity, 1996.)
An enormous amount of philosophical and technical thought about information comes from biology. This is not so surprising given the importance of the concept of information to genetics and DNA science. Inherited traits from one generation to the next of phenotypes (organisms) are described in terms of information. So is what is referred to as the central dogma of molecular biology: that information cannot go from the phenotype (the developed body) to the genotype (the gene/DNA.) In other words, if I cut my hand it will not mean that any child conceived by me in the future will have the same cut on their hand. More recently the central dogma has come under challenge from the field of epigenetics. In epigenetics, other things in addition to the gene – the DNA itself – are thought to contribute heritable information or information that is passed from one generation to the next. This can include processes within the cytoplasm of the cell, or even things in the organisms environment like the structure of nests in which young are reared. Still – it is often information transmission that is of interest.
At least since Crick and Watson’s discovery of the double helix structure of DNA in 1971, biologists and philosophers of biology have been contemplating and arguing about the nature of information and information transfer in DNA and biosynthetic processes. Biosynthetic processes are processes in which smaller molecules are combined to form more complex molecules that have some more complex function (processes involving such things as the manufacture of protein and other biological structures from genetic material.) Such processes are frequently described in terms of information.
Codes, encoding, transmission, and even information compression have been discussed as real in the processes of genetic material.
This all raises a question, however. We saw in the previous section that there are many conceptions of information. So which is the right one for biology? Molecular bioscientists and philosophers of biology are still trying to figure that out. There are even arguments about whether genetic information is semantic or not – if it has meaning and if so in what way (See recent work by Nicholas Shea on what he calls Infotel semantics. The idea is that the meaning of genetic information is determined by its function.) Some philosophers of biology even have what is known as an eliminative conception of information in biology: they eliminate it from the discussion completely or partly as a useless metaphor that is confusing and does not explain anything real (See Griffiths, Paul E. Genetic Information – A Metaphor in Search of a Theory http://philsci-archive.pitt.edu/89/1/Genetic_Information_etc.pdf.
Are There Informational Laws in Genome Evolution and the Evolution of Protein Synthesis?
This entire area of the nature of information in molecular bioscience is complex and keenly debated. However, in this two part series I am interested in a very specific part of the debate – one that is perhaps the most exciting and relevant to philosophy in general and not a little evolutionary science today. It involves the question of how protein synthesis evolved by natural selection. The process of protein synthesis is an incredibly complex biosynthetic process that has only recently come to be well understood. The complexity of the processes of protein folding and gene splicing meant that the details of these processes were wholly mysterious up until recently. How such processes came to evolve naturally to their current state is an even more challenging mystery.
Above is an artist’s representation of the proces of protein synthesis from DNA via processes of DNA transcription and translation into a chain of amino acids and finally into a folded protein. The process is staggeringly complex, with only the most basic fundamental steps represented here. Molecular bioscientists usually take it for granted that there is information transmitted form the DNA to the protein. A much larger question, however, is how the information of the entire process and the structures involved in it came to be as it is by evolutionary processes. Eugene V. Koonin
has proposed that “Although a complete physical theory of evolutionary biology is inconceivable, the universals of genome evolution might qualify as “laws of evolutionary genomics” in the same sense “law” is understood in modern physics.
) The details of this theory involve the laws being expressed largely as statistical and informational.