It must be admitted that you have, Dawkins, moments when strident utterings as those cited in previous are easy to forget, because you come forth as thoughtful - for the moment. If you really were, you might have answered some of my criticisms in the past. Here I am making a new one, anyway.*
Redundancy is any part of a message that is not informative, either because the recipient already knows it (is not surprised by it) or because it duplicates other parts of the message. In the sentence “Rover is a poodle dog”, the word “dog” is redundant because “poodle” already tells us that Rover is a dog. An economical telegram would omit it, thereby increasing the informative proportion of the message. “Arr JFK Fri pm pls mt BA Cncrd flt” carries the same information as the much longer, but more redundant, “I’ll be arriving at John F Kennedy airport on Friday evening; please meet the British Airways Concorde flight”. Obviously the brief, telegraphic message is cheaper to send (although the recipient may have to work harder to decipher it — redundancy has its virtues if we forget economics). Shannon wanted to find a mathematical way to capture the idea that any message could be broken into the information (which is worth paying for), the redundancy (which can, with economic advantage, be deleted from the message because, in effect, it can be reconstructed by the recipient) and the noise (which is just random rubbish).
Dawkins, information and redundancy are NOT two different things.
I am in many subjects self taught. In Latin that is NOT the case. The Latin professor who gave me a lesson about redundancy was himself an atheist. You can hardly accuse him for praising redundancy out of a bias in favour of Christianity.
Here is your example of non-redundant versus redundant version of same message:
“Arr JFK Fri pm pls mt BA Cncrd flt”
“I’ll be arriving at John F Kennedy airport on Friday evening; please meet the British Airways Concorde flight”.
Let us subject BOTH versions to extremely bad conditions of transmission, first, namely three letters kept, three letters lost, and so on:
“Arr ... Fri ... ls m... Cnc...lt”
“I’ll ...rri...g at ...n F K...edy ...por... Fri... eve...g; ple... mee...e Bri...h Ai...ys C...ord...ight”.
First version would set you asking yourself where Concordes will be landing on Friday. Supposing you even understood it was about the person. And when. Second version, with much more redundancy will let you restitute the missing letters pretty easily. Even more. Supposing the receiver were replaced by a child or a stranger (as in real foreigner) who did not know BA could mean British Airways or Concord be a plane ...?
Now, speaking of Concorde and of Concord, in Latin - to resume my professor's lesson - precisely concord (without an -e as to English version of the word) is a redundancy feature. Does that mean it is useless? No.
French has considerably reduced audible concord by muting final e:s and s:s. And t:s and z:s too. The price is that it is much more often ambiguous to the ear.
There is actually a gliding scale between noise and redundancy even apart of this ambiguous relation between redundancy and essentials. There is also the question of when noise is part of redundancy.
In Swedish, a word like "hvarje" can have its final unaccented syllable pronounced either like French -eu or like French -é, except for the latter carrying the accent in French. I can hear the difference, but to me it is a difference of noise. Not of information.
In French there are so many words where "eu" or "é" before the tone (thus also unaccented, like in Swedish) is part of the constitution of a word. Not that you would find both "renard" and "rénard", but if meaning vulpes you say "rénard" instead of "renard", you will be marked as a foreigner. Other example why "noise" and "redundance" overlap: the background for Welsh "rhydd". Now, I do not happen to think one can prove there was an Indo-European proto-language. But one can prove with somewhat greater probability word by word there were proto-words. There was one for "rhydd" and "free" probably and "Indo-European" state of it would be reconstructible (and is reconstructed by those believing proto-language) as *prios.
This *prios would give frî in Nordic and Old and Middle High German and frei in Modern High German. For English I presume there was a stage in which *prios was reduced to *priws to give Anglo Saxon freo, of which we get free. That much dialectal variation in pre-Germanic dialects would not have impeded both dialects going through the Germanic sound changes.
Now, for *prios > *priws I presume that there was a stage in which a mere noise - the slight w in *priwos - became seen as constitutive.
For Welsh we have a certainty that *prios or later *frios, *rios had another noise, yod in *priyos, which became constitutive too. It became the eth sound which Welsh spells with double d in "rhydd". Precisely as *ios ending, like -ius in Latin, for same reason comes out as -edd in Welsh, often enough. In all these cases the original -os part of word or ending is of course lost.
As the word is found only in Germanic and Welsh or perhaps other Celtic as well as far as I know - someone knowing Sanscrit or Hittite may correct me in case I am wrong - we actually do not know if there was ever a stage in which it began with a p. Real indo-european words with real indo-european p, meaning words found in many families and always or usually with a a sound that can be reduced to original p plus preservation or plausible changes thereof do show lack of such consonant in Celtic (rhydd has no consonant previous to rh) but initial f in Germanic (frî, freo, frei ...). But if the word was only there in Celtic and Germanic (which is possible, unless a Sanskritist or expert on Nesili or sth says otherwise, or unless prijateo / przyjaciel as in friend in Croatian, Polish etc is related) we cannot confidently say it had a p. An original f would have been lost in Celtic as an original p was lost via f and h. And in Germanic it would have been preserved, precisely as f from p was preserved.
That is however an aside onto the other question of whether there was an Indo-European proto-language. The thing I wanted to illustrate was that *prios or *frios, if there was such a word, would first get a noise between the vowels - either yod as continuing previous vowel or waw as anticipating the following one would be better than a hacked h - and then the noise would be considered as part of the word and develop to other sounds or sound combinations. In Semitic languages short vowels seem to be sometimes considered as noise in relation to basic word meaning - while carrying a very subsidiary meaning of their own. In most Indo-European ones (not perhaps Sanscrit, totally) they are considered part of the word. This is the distinction between phonetic and phonematic description of a language.
So, a linguist can at once say that it is not an easy task to distinguish noise from meaningful parts of a message. Even more, noise is part of the redundancy and redundancy is part of the message. A kind of safety back up for the information.
And as money does not count for God when He creates us, His use of redundancy or even "apparent noise" in our genome is not quite an argument for information being random. Dawkins has also said himself that genomes are more like recipes than blue prints. This means that there must be markers for "let it take some time, like fifty cell divisions or so, and then ..." as well as markers for "let's make a proteine".
Thoughtful as you were, Dawkins, this time, you were not accurate in your argumentation.
I can give some creds to Royal Truman for linking to Dawkins - though the essay where I discoered him and indirectly the link in another one was a dense and very technical one for someone a linguist rather than a computer scientist - but also great credits for one of his points in it:**
Theorem 2. A CIS can be used to achieve a mental goal.
In the absence of wilful input, the organization of matter can be explained by deterministic laws of nature and statistical principles of randomness. Mental processes, however, are not controlled deterministically or by randomness, and include: making choices; seeking to understand; and developing a strategy.
Suppose you are learning German, and are reflecting on what Unsinn might mean. The intention to translate is surely not deterministic, nor explained by randomness. Perhaps the intention is stored temporarily (physically?) in the brain, with which an immaterial you interacts almost instantly. You know what you wish to do and can easily communicate this to others. This intention is converted somehow into a physical search through the data stored in your neurons. This requires very special mental equipment, since a multitude of kinds of searches are possible: for a discrete telephone number; for how a face looks; for a melody. The list is near endless. In this case we’re searching for a concept which we believe reflects the meaning of Unsinn.
The concept we seek to translate must be encoded in some manner, and the searches directed efficiently. Suitable data must somehow be extracted from the neurons, requiring further mental machinery, and the results must be encoded and transferred somewhere for the mind to evaluate. Another tool then compares a candidate English word, the associations of which get compared with those of Unsinn. It is absurd to argue neurotransmitter concentrations or electrical signals are being ‘compared’ across billions of neurons. The logical processing must involve some kind of compression and high-performance language. Eventually the mind decides whether a potential translation of Unsinn, like nonsense, is correct or not.
All the resources involved in mental processes like these are part of a CIS, and some are not physical. Various resources narrow the range of possibilities, including when the translation is to occur and where.
It is about the same point I have been making of an abacus not understanding itself what the mathematician using it is understanding. I e, the necessity of a substance distinct from matter and mindless energy, which can be conveniently called spirit, in order to explain our experience. This was also brought out by a commenter:
Lachlan W., Australia, 23 May 2014
Congratulations to Royal Truman for his groundbreaking CIS theory. Although I believe the word 'spirit' is not mentioned in these articles, the implication is clearly there that this is where information, and the systems that contain and control information, come from. His breathtakingly simple phrase; "And our world is itself embedded in higher CISs as part of ultimate goals.", was a truly excellent way to sum up why there are any information systems anyway. I hope he will continue to use the Spirit God has given him to bring to light the interaction between the spiritual and physical/material worlds.
Hans Georg Lundahl
Bpi, Georges Pompidou
24 / V / 2014
* Australian Skeptics : The Information Challenge
By Richard Dawkins
** CMI : Information Theory—part 4: fundamental theorems of Coded Information Systems Theory
by Royal Truman