[This is #2 in a series of posts exploring information: what is it, where does it come from, and what does it teach us about ourselves and the reality in which we live. – #1 can be found here]
When Bach sat to write his great works, he encoded sequences of tones in time, he put into physical form on a page a set of instructions for reproducing melody. Such instructions are a form of information; they can be encoded using energy and matter, but they neither arise from them nor depend on them. The nature of the chemical bonding of ink and paper does not determine the content of a piece of music represented using them, just as the physics of electrical transmission and the propagation of sound do not determine the melody of the music these phenomena can be used to carry. It is in fact vital that the ink and paper do not automatically configure themselves into a fixed pattern of symbols. It is the flexibility of the ways ink and paper can combine that enables them to store such a virtually unlimited variety of informational content. To take the example of Bach’s Sonata, the ink and paper, the electricity and the telephone wires, the air of the Royal Albert Hall: all of these can be considered blank slates, waiting for an input of information that does not inherently exist in them. To return to our previous Figure, the information input which courses through all these physical forms actually arises originally in a mind.
All the transformations from step 1 to 29 are like domino falls in a sequence, all flowing from an initial information input (1) from Bach’s imagination. Thus, even leaving aside temporarily the deep question of what a ‘mind’ is – we see that the kind of information contained in Bach’s music or Shakespeare’s sonnets appears to arise somewhere prior to its expression in the tangible stuff of the physical world. The information content is determined at step 1, before it ever finds a place in the subsequent physical representations.
But, we might argue, this priority of information over matter and energy does not always seem to be the case. For there are kinds of information that do appear to arise from the basic structure of physical reality.
We could, for instance, represent the chemical bonds of a chemical compound in symbolic form, or encode as a dataset the acoustic pattern of waves crashing onto a beach. We could point radio receivers towards the skies and record patterns of electromagnetic radiation arriving from far-off events. Nobody would argue that these ‘data’ do not contain information; indeed, we could collect enough of these data to fill countless hard drives should we so choose. And, unlike the information found in novels and songs and poems and family conversation – information which arises somehow in minds – the information these environmental data contain does seem to ‘arise’ directly from the properties of matter and energy.
So does this mean we are wrong to say that information is transcendent to matter and energy? After all, the patterns in crystals, the periodicity of waves, the spectra of cosmic radiation arriving on earth, do appear to be information derived directly from the structures of physical things.
But is all information is created equal? Is there no qualitative or quantitative difference between the kind of information in the bonding structure of a crystal, and the information in Bach’s Sonata in G minor or in the soliloquies of Shakespeare’s Hamlet, or in the text of the Bible?
On the face of it, there is in fact a vast qualitative difference between environmental information (the data which spills out of the motion and structure of physical objects), and information of the kind contained in freely-ordered communication between the minds of intelligent beings.
We might illustrate it by analogy. Consider the difference between the information content of two example sheets of paper. One page contains an accidental splatter of ink, the other page contains the opening bars of Bach’s Sonata No. 1. Now, we would say that both pages contain a kind of pattern, and further say that patterns certainly contain information. Furthermore, both patterns are clearly expressed in the form of ink on paper.
But the information represented on the two pages is not the same. Qualitatively speaking at least, the information on them is of a very different kind.
We would find the same kind of qualitative difference between the information contained in a 60 minute recording of waves crashing against a beach, and the information contained in a 60 minute audiobook of Shakespeare’s sonnets. Both recordings would contain information, both could be stored in a binary format on a hard drive, both could have the same length when played back as audio, but the information content of the two recordings would be an utterly different kind of thing.
So how can this intuitively sensed difference be expressed and accounted for?
In trying to define what distinguishes the ink splatter from the sheet music, or the sonnet from the sound of waves, the concept of language is important. Whereas the shape of an accidental ink splatter and the sound of waves are what we might call ‘unmediated’ (in that they flow directly from dynamic physical processes, the interaction of energy and matter, without the involvement of human minds), the music and the sonnet arise from imaginative processes in minds, and are subsequently encoded using a language system.
The kind of information compiled by minds and expressed through language is sometimes called ‘semantic’ information; the word ‘semantic’ simply meaning to indicate by a sign. Semantic information is the kind of information we represent and communicate using words and symbols in the form of conversations, poems, songs and novels. It is the kind of information encoded in the hieroglyphs of the Rosetta Stone, the words of Dante’s Inferno and the notation symbols on the musical score for Handel’s Messiah. Indeed, whenever intelligent beings use systems of signs – languages – to express and communicate imaginative concepts – whether a shopping list or a novel – they engage in the sharing of semantic information.
Signs and symbols – words, numbers, mathematical operators, and musical notes – are useful because unlike cognitive sentiments or concepts in themselves, symbols can be represented using matter and energy, making them communicable to other intelligent beings. For instance, the information content of the feeling of anger in a mind cannot in itself be expressed. But by mapping a recognisable sign onto it (say the facial pattern of a scowl, or the words “I am angry”, or even simply a roar of aggression), the sign can then be expressed through matter and energy; whether as a visible facial expression, or as letters inked on a page, or as sound waves transmitted to the hearer.
We started this series by considering a particular form of content expressible through language: namely music. But of course music is only one example of the kind of information expressible through language. We could just as easily substitute Shakespeare for Bach, the English language for Western music notation, a manuscript of a play for the musical score, sequences of words for sequences of notes, an actor’s vocal chords for the strings of the violin, and the theatre for the auditorium. The fundamental scenario from the start of this piece would be the same, it is one of semantic information being transferred from mind to mind between numerous intermediate representations in the realm of energy and matter.
As we have seen, semantic information is not itself reducible to the matter or energy used to represent it. In other words, we cannot account for the content of songs, novels, poems and the like by looking at the physical form of their representation. Knowledge of the bonding of ink and paper will not allow you to predict the semantic information of a printed novel. Knowledge of the laws of electrical transmission will not enable you to predict the semantic content of an email. These forms are merely receptacles for pre-existing semantic information.
Information and language
This independence of the structure of semantic information from constraint by matter and energy is also true of constraint by language systems themselves. After all, knowledge of a given language or notation system does not enable us to predict the information content of imaginative works that can be encoded using it. For instance, a complete knowledge of the Western music system does not allow us to predict the information contained in the Sonata No. 1 in G minor. A knowledge of French grammar and vocabulary does not let us automatically produce the plays of Voltaire. Semantic content is not contained in the system itself. Languages are a receptacle not a source of information: nothing about the laws of a given language system renders ‘obvious’ the combinations of concepts and sentiments that may be expressed by it. The nihilism of Nietzsche and the good news of John’s Gospel can both be just as readily represented in English, but no amount of grammar or dictionary knowledge would make us more likely to ‘predict’ their vastly different information contents.
Thus information as an entity somehow precedes both the physical forms and the language systems that can be used to represent it, whether the language in question is Western music notation, French, computer code, sign language, binary, the nuclease system of DNA, or any other.
This means we might say that semantic information is not therefore an ‘emergent property’ of either matter, energy or language: in other words, it does not come from them, useful as they can be in representing or transmitting it.
Just as leaves thrown into the air reveal the pattern of the wind, matter, energy and language can make semantic information tangible. But just as the blowing of the wind does not depend on the leaves, so information does not depend on the various forms which reveal it. A note echoing around a concert hall may fade out, a manuscript may be destroyed by fire, languages may die out – individual expressions of information may come and go – but the information itself is transcendent and apparently immortal. Every despot who has ever tried to destroy a culture has faced this elusive nature of information. People may be slaughtered, towns may be razed, books may be burned, but knowledge, literature, tenets of faith and hope; these have a way of dancing above the frail and temporary vessels of matter and energy which may for a while transmit and bear their imprint.
The separability of information from stuff is not an intuitive principle for us to grasp. Quite naturally, we strongly associate information with matter and energy and because this is how it is made tangible to us. We hear the sound waves of the spoken word, we see a beam of visible light from the printed page, we carry around digital memory devices stuffed with binary representations of texts and music and films. But in all these cases, we must remember that the information itself is not dictated to any particular configuration of matter or energy.
We also quite naturally associate information with certain language or notation systems, equating the information content with the representation of it we most commonly encounter. When we think conversation, we think of our mother tongue, when we think music, we think staves and quavers and minims and crotchets. But as we have seen, information is no more tied to a particular system of signs and symbols than it is tied to a particular configuration of energy or matter.
In fact, we know this latter truth intuitively, for we encounter it every time we translate a text (i.e. a collection of semantic information) from one language to another; in other words, from one system of symbols to the next. Though most people become familiar with expressing linguistic content using a favoured set of grammar rules and a preferred vocabulary (i.e. our mother tongue), it is important to see that the information content of our speech itself does not depend on any specific language system. Thus a text such as the Bible can be translated into numerous different languages, to the symbol systems of Arabic, Mandarin, Sanskrit, or Cyrillic, yet the information content can be virtually unchanged across each representation.
In this post, we have explored differences in kinds of information, recognising that there is at least a qualitative difference between environmental and semantic information. Environmental information does appear to be innately tied up with how matter and energy interact, whereas semantic information is freely ordered, flowing from a mind.
We have seen how semantic information is a distinct entity from matter, energy and language systems. These forms can be used to express it, but because they are effectively secondary to it, they do not provide much assistance in answering questions of what information is and what it tells us about reality and ourselves.
To go deeper into these questions, we will have to get into the nuts and bolts of what information – both semantic and environmental – really is, and what its characteristics can tell us.