Hypothesis – my thoughts

(with borrowed bits!)

Inductive logic – building from facts to general principles

Logical deduction – starting from general principle, moving towards incorporating facts.

Analogy and common sense

Building a cohesive theory out of the background noise (if you want to be "first boy in the class to stick up his hand" you'll have to be fumbling about in the noise).

Identifying and dismantling misconceptions (parochial presumptions).

Thought experiments

Illustrative metaphors – after all we need some sort of metaphor to help understand everything. The earth doesn't think "shall I make my calculations from a sun centred solar system or an earth centred universe?" – it doesn't need to simplify things this way to "compute" its trajectory. No, this is a fop to the frailty of human visualisation of the process. This leads to the irrepressibility of teleologies. We must simply remain aware when we are using them and remain aware of their limitations – not try to wipe them from existence.

Long before detailed experimentation

Need for a "Department of Theoretical $$" – where the dominant process is thinking (see this article)

Identifying and deconstructing "religious beliefs" (religion derived from religare:– to tie up, tie back). Here, I am using "religious beliefs" to denote any beliefs and am not restricting these to just theistic beliefs.

The emphasis of tautologies – particularly where their current emphasis is poor

Knowledge is only what is scientifically verifiable – it is interesting to see how much opinion is presented as knowledge in any immunology text (eg, immune system primarily for fighting infection).

Hypothesis focuses attention on key issues. It suggests proportions. It points to a way forward. It offers (sometimes radically) short cuts to verifiable scientific progress.

"Stale paradigms" need to be recognised as such. Any assumption must be qualified and justified. Parochial presumptions.

Broad expansive front – focusing in

Focused specific front – expanding out

Both are justifiable approaches but remember that the second is closer to the "naive inductivism"

The legitimacy of any method of hypothesis lies in:

Hypothesis, in its early stages, has many riders, conditionals, ifs, maybes (it can make the language unpleasant)

A broad, expansive-front hypothesis is bound to invoke distrust. A focused, expanding hypothesis is much safer and more cautious. The broad hypothesis must be incautious, it must be highly speculative to start and focusing in to finish. The assessor of the hypothesis must be prepared to (temporarily) abandon cherished beliefs and consider the possible benefits of a changing (often inverted) perspective.

Certain basic rules can be set.

Can it be demonstrated that an alternative viewpoint might be legitimate? Is that viewpoint both logical and possible. If current fact cannot clearly dismantle this possibility then current perceptions must be regarded as potentially flawed, particularly should they be held as cherished scientific fact (eg, lymphocytes discriminate self from non-self, the immune system arose primarily to fight infection, lymphocytes attack pathogens).

Ultimately, rigid mathematical modelling of the process may be considered as the most definitive proof of a theory. But that can only come after the groundwork, unless it is confined to highly specific areas. And it may never be realistically achievable across a very broad hypothesis.

First hand and analogous experience can be valuable in reaching broad conclusions – remember the metaphors.

Bacon's idols – arbitrary, unfounded, superstitious assumptions (parochial presumptions).

Self evidently valid concepts – plain wrong or tautologies that need re–emphasis so we don't forget.

The bizarre world of "double-think" – instances need to be highlighted.

Making the workings of nature intelligible by organised and compact theories (metaphors again).

A theory will be of value if diverse facts can be hung onto it without radical shifts in emphasis. The "Rube Goldberg"/"Heath Robinson" syndrome must be recognised. Constrictive, conditional systems are unstable. Permissive, self righting systems are stable. A theory must allow that the latter is more likely. In a good theory, where the detail won't fit, it should point out failures of the detailed interpretation of general principles which, when corrected by minor, non–conditional shifts in the principle, lead to a more satisfying and complete hypothesis. This hypothesis, for me, feels like a 5000 piece jigsaw falling together. There are bits that are untidy and a few pieces in the wrong place but the overall picture and proportions are rapidly becoming clear. Simple and frugal – a paucity of complication. Beauty is probably a pointer to its authority.

"Resist generalising early on" – I thoroughly disagree!! Resist the temptation to jump to rigid conclusions early on is absolutely right. I did it but my personality allowed me to ignore them eventually – but not after much embarrassment. The broad hypothesiser should learn, from the start, that conjecture is good, rigidity is bad. Expect to abandon the majority as trash. Allow it to evolve in large see-saw sweeps – much like the process of evolution itself – try a new schema and allow it to settle in – where it won't, modify it. In hypothesis, nothing can be considered scientific fact until it is validated by focused experimentation from many independent centres (falsification and reductionism at its most laudable). The more focused a hypothesis can be, the easier it will be to validate – and so the "safer" it is to support. The broad front hypothesis is a much more invidious beast. The rewards are higher if it proves well founded but the debit for any protagonists is that they may lose credibility along the way. Few people in any science these days are in the position to take the very broadest view of a subject. But such a broad hypothesis can be tested by focal specialists – but they must realise their aim is not to destroy specifics in the hypothesis – because a generaliser cannot, by definition, have sufficient knowledge of a specialist field – but to see if their expertise can mould that hypothesis into their specialist arena in such a way that it enhances their own perceptions. When this fails (it is potentially falsified), then the main hypothesis can be said to be wanting.

In searching out a solution, look for an overall pattern. Keep your mind open for themes – they are frequently repeated. Assume everything will fit into a grand design. So all disease, eg, will reflect the underlying principle. Look for general patterns. Speculate broadly on an explanation. Do not believe that explanation. Throw it open to falsification. Ultimately, bits of the explanation will stick. They won't be falsifiable. These can enter the next round of broad speculation. Listen to your own writing – my experience is that the writing has constantly been more mature and clear in concept than the mind of the writer.

A good broad hypothesis will act like a framework. The specific detail will be altered but the grand outline, particularly the proportions and the weight to be attributed to each part, will become clearly apparent. It will act as a scaffold on which to peg each relevant part. In that respect, this hypothesis is powerful. From experience, I know that I will be able to take any disease or any new knowledge and peg it on the scaffold without substantial threat to the overall integrity. Local structure will almost always need to be refined. So far, I have toyed with but not proceeded with exploring atheromatous diseases, HIV infection (eg) and pegging them onto this structure. From experience, I know that it can be done. It will take some reading and investigation but the process will be worthwhile because it will bring greater understanding to these subjects. But I cannot be omniscient and this is, after all, a spare time hobby – not a full time job.

Predictive success is its final goal. I find Horder's quote interesting here because he accepts that there is the probability that a grand design can be found without the elements being clearly definable by specific focused proof and it is largely a matter of perspective – rather like those pictures of a young woman that can be re-interpreted as an old woman

Simplicity and parsimony of explanation – seeding a small number of central ideas and showing how they can accommodate the facts. Beauty is a pointer (no more than that). Don't confuse simplicity of function (or explanation) with simplicity of principle. The function (and explanation) of the mammalian immune system is highly complex but the principle of "tissue homeostasis" is simple. It gives inflammation a central role to remove offending debris and restore tissue structure (by regeneration). The amplification and memorisation of previous inflammatory events (thanks to lymphocytes) adds much complexity. But, the underlying principle remains exquisitely simple.

Lymphocytes – "self evident" that they discriminate non-self from self –– think the unthinkable!!

Simultaneously insist on rigorous definitive confirmation whilst formulating loose, far reaching speculations – testing, falsifying, focusing in, confirming.

Example of the latter: God making the world (Genesis style). He may have deliberately fooled us and made it look like natural evolution to test our faith. So what – he made such a good job of this that we may as well accept it and get on with things as they (practically) seem. Indeed, if he did, then the "truth" would be a bankrupt resource to help us manipulate the system. Going with the "apparent" is then more fruitful as a tool for technological manipulation. So models could be more important than the truth.

Likely winners – front runners – get to recognise them.

Time!! – mull over the hypothesis time and time again – the bits that stick could well be true

The biological sciences have moved into a state of "contented consolidation". The feeling is that everything of general importance has been worked out fairly well and what is left is to fill in the detail – paint in the empty sectors – fill in the blanks. There is now a conviction in biology that the only things worth presenting are detailed validations of increasingly more detailed dissection (reductionism). Science does, indeed, demand that rigorous dissection and proof of the detailed components is carried through methodically. That is the goal. But there is more than one way of gaining that goal. The first is to plod step by step from one point to the next, gradually expanding the horizons until the overall grand design becomes clear. The other is to try and drag up that grand design out of the background noise. The first is safe and will get there in the end. The second is fraught with danger but it could lead to a stampede towards the finishing goal and shortcut much tedious and methodical repetition of none critical work. The "stampede" can then be channelled along the critical paths, up turning the most productive stones, predicting sections of the grand design that, when authenticated, support the hypothesis along the way. In pulling out grand design from the background noise it is inevitable that the picture will be fuzzy. The detail will be patchy even when a vague pattern is emerging.

In hypothesis there lies a trap. When a part of the hypothesis predicts correctly it tempts us to accept other and unsubstantiated areas of the hypothesis. If no one recognises that this is what is happening these unsubstantiated points become enshrined in the holy grail of accepted dogma. The answer is – never believe any hypothesis wholeheartedly. It might make sense now but assume and accept that it is only a temporary guide to a better perspective. Be ready to overturn and usurp it at a moments notice for anything that is clearly better. Broad paradigms are like religious beliefs – we believe they must be protected at all costs. Science has taught us one thing through the ages – never trust that dogma is right. Suspect the philosophy of those who hold beliefs with absolute conviction; the true thinker ought to believe things with strong, but not absolute, conviction.. Defending old paradigms is like unresolved grief for the dead – don't look to maintain rigid reverement of the dead but look for the signs amid seeding of their continuing influence among the new and the maturing living.

Good old mother earth doesn't sit thinking "shall I make my calculations for where I'll go next from the perspective of me being at the centre of the universe or the sun being at the centre of the solar system?" It doesn't need this simplification to get on with its task. We, on the other hand, poor simple, parochial beasts, have got to find some metaphorical analogy that lets us appreciate better and calculate more easily what is going on. So "the truth" is an illusion of our own simple-mindedness. There are just better and more and more facilitative models.

Down in the noise – looking for patterns – like Visual Evoked Potentials. Repetition builds them up out of the noise.

Progresses by evolution. Think of wild ideas – try and base them around certain anchorage points. They can be as wild as you like. Now go through the process of falsification-rejection-modification and do this cycle continuously. Never be content that it is 100% right. Pick up the clues (patterns) and look for these. Listen to the (your) writing!

Theory structure should stay loose. Could be regarded like a useful "mnemonic" that allows us to hang our facts in such a way that we can learn and use them better (learn, apply, extrapolate). No one "mnemonic" should be regarded as having more "truth" than any other. It should be judged on its success (as defined above). The success of a theory is not, therefore, its intrinsic "truth" or authenticity but resides in its practical value. The evolutionary principle of progressive approximation will ensure that the "most useful analogy" (and indeed the "truth") is, indeed, subtended to. "Truth" does not matter in a practical (powerful technology) sense provided the system gets results.

A grand unified theory has more attraction than multiple variegated theorems. Hence even for understanding inner planetary motion, Einsteinian physics has more power and attraction than Newtonian (even though Newtonian is a good approximation). Einsteinian physics incorporates Newtonian physics where it appears to predict correctly and highlights why it doesn't work where it doesn't – so it's a more attractive theory. Similarly, superstring theory and multi-dimensional mathematics are now being looked to for a deeper explanation still.

So, until a theory comes along that offers a better "hanger" for your facts, choose the most productive "mnemonic". If you choose a model that lets you cure 99 out of 100 cancers and an alternative model only lets you cure 10 out of 100 it doesn't matter which is closer to the "truth" in any practical sense. Leave pedanticism to empire builders and get on with the power.

Modelling, mathematical is the ultimate; but it must start with a valid and sufficiently compelete set of parameters.

Complexity, simplification. Reduction to simple concept and analogy. Plain English. The success of a theory shouldn't be judged solely on its modelling accuracy (mathematically, these could be complex, long winded, ravenously-hungry-for–computer-time computations). Indeed, it is likely that, given a knowledge of the appropriate mathematical tools and a powerful enough computer, we could simple feed in the basic rules and spit out the answers. In the meantime, while we wait for the computing power, we have to use an "Occam's razor" approach. Short cuts are acceptable. Their value is in manipulative success. Leave purism to the purist side of your nature and see and accept it for what it is. Don't let it stifle practical progress.

Why is science powerful? Because it delivers. Science has been more successful in treating tuberculosis than faith healing (we wouldn't have needed medical science if faith healing had been 100% applicable and successful). The goal of "truth" might be a fun intellectual game but it is not a practical tool except insofar as the working "mnemonic" will probably subtend to the "truth" anyway.

Make us ask the right questions

Solve practical riddles

Divided into

"Truth" is a state to be welcomed but it should not be slavishly sought. It should be allowed to assert itself gradually. Stay loose. The main danger in seeking the "truth" is that theories become "religious beliefs" that become a heresy to challenge. Likewise, in the scrum to swap ship and take up a new "truth" the danger is that this new "truth" will be regarded as a new "God". The history of science tells us one thing – current paradigms are likely to be replaced and often in a dramatic way. Stay loose in moulding your perspectives but tight in the level of proof required! Whatever, the "Truth" (perhaps "Reality" is a better word) is probably rather like an asymptote – approachable but never realisable.

The economy of explanation. Getting through to a root principle that can simply, without ad hoc extensions, be applied across the board. Such an explanation/principle has intrinsic beauty and simplicity. It is, by my reckoning, more likely to be right. Explanations that require the existence of many non–interlaced principles are, in a simple statistical way, less likely to "coincide" obediently and conveniently. My guess is that a search for such "economy" has been the main driving force in my "method".

Parochial presumptions – break away from your parish and don't presume anything.

You may not know what "the promised land" is going to look like when you first set out. But, as it emerges from the mists of obscurity (the background noise), it becomes increasingly integrated, logical and elegant. Then its veracity and extensions become more obvious. A new and deeper horizon becomes increasingly recognisable.

I am going to resort to the arrogance of believing that my "method" has come up trumps and try and sum up what is the overriding theme/method. Most modern research is dissecting and differentiative. (I think that) I have constantly used a technique that is joining and integretative.

Lastly, remember that everyone is probably a little or a lot right but some may be right for the wrong reasoning. Remember "The blind men and the elephant" .

Some (preliminary) thoughts on the logic of language

One thing that has become evident about nature – and it is particularly noticeable when viewed from a perspective based on physics – is that it is eminently describable as a logical construct, particularly a mathematical construct. The whole foundation of (evolving) science puts implicit – though not always explicit – faith in this trend. Thus far, there is no real need to distrust the assumption that even the most abstruse of subjects (like, for example, psychology) will at some stage in the future become more and more describable and analysable using the fundamental language of basic physics and mathematics. And it is worth remembering that the grosser manifestations of our physical universe can be conjured up, by extrapolation, out of quantal-stochastic-probabilistic-statistical descriptions of existence.

Now, this is a wonderful property of science. As Mark Twain put it "There is something fascinating about science. One gets such wholesale returns of conjecture out of such a trifling investment of fact." And this leads to a surprising philosophical conundrum: why is our universe proving to be such a logical construct? It could be inferred that it has the appearance of having been put together by a mathematician. However, this may be missing the point. Does our universe emerge from the fabric of logic and maths or does logic and maths emerge from the fabric of our universe? The quantal nature of "substance" (at Planck distances) may make this mathematical derivation inevitable. Quantal statistics may be at its core.

Logical thought experiment, based on analogy with other elements of the observed world, can generate insights that a purely empirical (statistical) approach will rarely reveal because the latter leaves grand principle submerged in a sea (even an ocean) of data. However, both approaches – multiple empirical observations on the one hand and a holistic perspective on the other – can, when combined, lead to better insight. Quantal-statistical-probablistic analyses rule at Planck distances and thought experiments have proven to be fruitful in relativistic physics. And, we suspect, the resolution of the conflict between these two might, one day, open a window on a much deeper understanding of the nature of our existence (our poignant, parochial concern).

Language itself is a form of logic; and it is constrained by logic. However, rather than being analogous to a simple Boolean logic it is analogous to a more complex language like "C". In social language we can conjure up a three lettered word that encompasses an expansive set of elemental meanings. For example, the verb fly; this encompasses forward motion, ascending, descending, acceleration, deceleration, aerofoil physics, energy utilisation and loads more – all summed up into a pithy three lettered shorthand. Try expressing this in Boolean terms! But such shorthand also exposes an Achilles heel in the logic of language because, unless we tend the precision with which we use our nouns and verbs (these especially), the logical power of social language will be eroded.

Mathematics is a language too. The parallels are obvious. Operators like "plus" "minus" "times" "equals" are equivalent to spoken language's verbs. The x and y letters of algebra are equivalent to spoken language's nouns (w + x = y) and the numbers that qualify them are equivalent to our adjectives (2w + 5x = 3y ; qualifiers). So, language itself is a sort of logical construct. In science (which aspires to being exact) definitions and meanings become far more important. Here there is little leeway to allow any hijacking of terms into new or loose meanings. In social language, an extant example of such a high-jacking is the word gay. The original word (exclusively an adjective) indicated a combination of being light hearted, carefree, outgoing, mirthful, colourful. In this sense, I am thrilled by the moments when the gay side of my character is dominant and I categorically refuse to abandon its original meaning. Now it has been largely, if not irrevocably, hijacked into a mixed adjective/noun that indicates a person who engages in homosexual relationships. Our younger generations are often unaware of any other meaning. My hobby-horse crusade is the way immunologists and microbiologists have high-jacked the word pathogen – "an agent that damages body tissues" – and distorted it into being a legitimate shorthand and synonym for a "commonly pathogenic (micro-)organism" (an organism capable of damaging tissues).

In social language it is inevitable that terms acquire fresh meanings with the passage of time and that they are even high-jacked into new meanings and new functions (eg, "gay" adjective to "gay" noun). However, in science it is far more important – if we want to preserve the integrity of this language's logic and precision – that we guard against and repair weaknesses that reduce its precision and the accuracy with which it can be extrapolated. Definitional precision is all important. It is inevitable that adjustments will need to be made to our terminology as more powerful perspectives emerge. However, it is unforgivable that our terminology should be carelessly distorted by a neglect of precision. Etymologies in science are far more important, I submit, than in social language though, even there, we might be wise to resist the erosion of precision. For example, in football-speak, "he played fantastic" is now approaching legitimate use.

So what is the take home message – where am I leading? Well, it is to point out that, by carefully applying language to an observed process, the language itself can reveal the logical structure of that process. A careful description using precise language may be just as successful and productive, in exposing new insights, as a series of laboratory experiments. That does not imply that we can do without our laboratories and experiments for they, themselves, have spawned a large body of appropriate logic (language) with themes that have proven common to a whole spectrum of problems. I think this is what I noticed when I suggested "Listen to your writing" (see above). Appropriate insights emerge almost autonomously whenever language is carefully chosen to describe the function of a complex process.

The cycle of progress

We need to consider a "conceptual reboot" whenever things no longer seem to gel properly; religiously hanging onto old concepts may lead us up blind alleys.

(I would need to review where I have borrowed the bits – possibly some from Chalmers, certainly also from other sources.)