Metaphors in science
It is interesting to read of the increasing interest that is being generated around metaphors in science. It is, of course, almost impossible to formulate any easy understanding of the natural world without resorting to metaphors in some way or another. Even complex mathematical descriptions can rely on a few core linguistic metaphors. In the biological world, metaphors have a further complication in that they frequently become teleologies as well. A teleology is an explanation that accords purpose to some attribute. So, the eye is designed to be some sort of video camera; the immune system is designed to defend the body against foreign organisms; hair is designed to insulate the body; feathers are designed to help a bird fly. Now, although all these attributes can be seen to be an extant advantage that has followed their chance evolution, there was no "purpose" in their emergence. Indeed, in many instances (eg, feathers) the origins of their chance evolution may have conferred an advantage of a quite different "purpose" but happened, fortuitously, to promote a new advantageous adaptation. Apparent "purpose" arises – as an emergent property – from the evolutionary survival of a minute fraction of multitudinous different mutational changes. Whenever we let ourselves get wafted away on a teleology, we must always remember to go back and weed it out once we can envisage the sequence of chance events that led up to that accumulation of changes that eventually appears to have conferred "purpose" on some evolvent attribute.
However, there has also been an intellectual snobbery that tends to "throw the baby out with the bath water". Emerging explanations (or metaphors) are the life blood of scientific revolution, even though consolidative science needs (eventually) to iron these out of existence. There has been a great trend to condemn metaphor and teleology, "out of hand", as unscientific; so much so in biomedicine that the literature generated in scientific journals has become increasingly sterile and unadventurous. Hypothesis and speculation have become increasingly regarded as heretical and corrupting.
But, look at any text book of immunology (for example) and you will find metaphor and teleology in copious profusion: surveillance, self/nonself discrimination, tolerance, repertoire, signal, proliferate, engulf, protect, digest, release, secrete, trigger, drain, encounter, organise, architecture, migrate, recognise, bind, dispose, generate, initiate, destroy, recruit, kill, activate, repair, trap and carry are just a few. Each of these is a word that has a "day to day" meaning but is now used, analogously, to enhance understanding of the workings of the immune system. Inevitably, these metaphors/teleologies are tailored to the reigning paradigm. In a paradigm shift much of this profuse metaphorical language will shift in emphasis or even change radically.
The religious fervour that scientists have for particular paradigms is largely entangled in these metaphorical analogies – and that is an anachronism. We must recognise that this dependence on metaphor is both a strength (helping us to understand – at a certain point in time) and a millstone (leaving us stranded up blind alleys and generating Bacon's idols).
We have to adapt our language (that has primarily evolved for social interaction and a description of our immediate world) to help us to understand a strange new "world". This is inescapable. So, we must simply remain aware of its limitations, its dangers and be prepared to recognise it wherever it is employed.
I think it is an interesting feature of metaphors in science that they will often appear stealthily then take on a life of their own; they sneak in without anyone complaining (perhaps even noticing). They may initially be used as a descriptive technique to bring common sense to bear on a discipline. This enables it to be described and understood in a relatively uncomplicated way. But once established such metaphors attract potent advocates and protectors. These people often adopt them, lock stock and barrel, as literal "truths" rather than useful analogies. Although their meanings have often shifted subtly, their authority become unchallengeable without first committing an act of heresy. Metaphors should always remain malleable and their validity should be frequently reviewed; and they must be acknowledged so that we can stay on our guard.
Another point about finding a valuable new metaphor (or set of metaphors) is that it (they) can open up a totally new perspective upon things. Not only does it help to appreciate better what may be happening but it also suggests a series of extensions that become what Kuhn called tautologies, "... statements of situations that could not conceivably have been otherwise". This leads to what I have alluded to in "listen to your writing" (see the hypothesis, my thoughts section).
Post script:
I recently came across this sentence "To understand modern physics and cosmology with real depth you (would) need to be a sophisticated mathematician." My suspicion, and I would bet on this, is that this statement will eventually prove to have been re-formable thus: "To investigate, manipulate and apply modern physics you need to be a sophisticated mathematician. To make sense of it – to understand it – you will probably also need to master the art of metaphor (whilst recognising the inherent dangers that accompany this)." [Quotation from p 76 Alan Wall, Myth, Metaphor and Science, 2009 Chester Academic Press]
This accords with Einstein's statement "You do not really understand something unless you can explain it to your grandmother." And that "something" is not about the depth or the nitty-gritty complex detail of a subject but about the grand principles that pervade it.
I came across this phrase today (170321) that, I think, demonstrates the merit of metaphorical thinking: "Innate immune cell swarming at sites of inflammation and infection". (Conference talk by Tim Lämmermann).
This is, to my mind, a good commentry on metaphor in science (its also in the "supportive" file). 
(Book – available on net - link lost - to find)