The Newsletter of the Philosophical Discussion Group
Of British Mensa

Number 103 : October 2000

17th September 2000 : Theo Todman

I posted the notice below on PDGlist in February this year, and received some response from a couple of list members - namely those reproduced below from Ernie Davidson and Bob Williams. However, the pieces people were asked to comment on were rather long and complex for an internet discussion group, and may do better when they appear in print. So here goes !


As PDGList is a bit quiet for once, I thought I'd post a couple of papers that came out of a Seminar that fellow list member Peter Mansell & I attended at King's College Cambridge (our Alma Mater) on Thursday evening (17th February 2000). The seminar was one of a series of "Provost's Seminars", so called because the current Provost on King's (Professor Patrick Bateson) was once invited to take part in a similar affair at Baliol, Oxford and thought they were such a good idea that he instituted them at King's.

The idea is to get together a large group of King's dons, research students, undergraduates & former members of the College (in about equal proportions) to debate some important issue. The format is for two dons to deliver papers pro and con the issue at hand (in our case "Does Science Tell the Truth ?"). Then, over dinner and for an hour thereafter, the seminar breaks into groups of about 10 people - there were 9 of these - to discuss a pre-set question (there were four different ones shared amongst the groups). Thereafter, the "findings" of each group are presented by a "rapporteur" (usually an undergraduate) in a plenary session and the speakers respond to questions either arising from the group sessions or generally. Finally, attendees may repair to the bar to continue discussions. The show started at 17:30 for coffee, dinner was a rather brief 45 minutes, and the formal session finished at 22:15. We carried on talking in the bar until gone 24:00. I have to say it was an immensely enjoyable and stimulating evening and the first of many I intend to attend.

The papers were presented by Professor Peter Lipton (Professor of the History & Philosophy of Science at Cambridge) and Dr. Martin Kusch (Lecturer in Philosophy at Cambridge & Director of Studies in Philosophy at King's), both of whom are Fellows of King's. Dr. Kusch's paper was delivered as a hand-out, which I have scanned in & appears verbatim below. Professor Lipton did not provide hand-outs, so the account below is due to me, I'm afraid, and cannot be relied upon for accuracy or completeness - I've tried not to "round it out" lest I put my own inadequate thoughts into Professor Lipton's mouth. I have to admit that I found Dr. Kusch's talk the more stimulating.

Questions included the one my group received "Wouldn't it be a miracle if a scientific theory that had made many successful predictions turned out to be fundamentally incorrect ?". A second was something like "All past scientific theories have proved false, so doesn't this make it inappropriate to suggest that current theories are true" ? A third had to do with definitions and categories of "truth", and I've completely forgotten the fourth.

Lots of issues and questions were raised in the discussion, but I've not recorded them - they were, in any case, far less coherent than the lectures and you've enough to be going on with !

Theo Todman

17th February 2000 : Professor Peter Lipton

... as garbled by Theo Todman ...


The purpose of the Philosophy of Science is to understand science better, not to make better scientists. This has led to most practising scientists having little time for the philosophy of science !

Professor Lipton supported the cause of Scientific Realism, that while the scientific programme is falibilist, and there have been many incorrect theories, yet science converges over time closer & closer to the truth.

What then, is truth ? This is a hard question. Scientific theories portray or model the world. Saying a scientific theory is true implies that it is an accurate model of the world.

Those who believe that science seeks the truth, who adopt the "truth view", would point to such things as the success of scientific predictions & the success of technology in support of their claim. Most, but not all, practising scientists adopt the truth view.

As a counter-example, Professor Lipton made a quotation from Stephen Hawking (not recorded by me, unfortunately!), the upshot of which was that there is no sense in which his theories are true or false; they are just empirically adequate (or not).

Alternatives to realism

One of the difficulties is that we're trying to determine causes from effects and working out what's there with a limited sensory apparatus.

Problems arise in two groups - how can we speak of truth "out there", and what reasons can we have for believing our hypotheses to be true ?

The eighteenth century Scottish philosopher David Hume believed that we can have no reason for the extrapolations involved in induction. Given any set of points, there are always infinitely many curves that fit them. Why chose simplicity rather than complexity ?

Even if we knew all the data and had a theory compatible with all the data, there will be other theories that equally account for the data. This is the problem of the under-determination of theory by data.

Some scientists, indeed most of those with any interest in the philosophy of science, are Popperians. Popper's theory is very simple - falsification. [There followed a brief and unsympathetic overview of Popper's Conjectures & Refutations thesis]. According to Popper, we can know a theory to be false with just one counter-example. However, there are problems with this idea. Firstly, it is not always easy to show that a theory is false. Secondly, this approach doesn't lead much beyond scepticism. On this theory, we are left with no reason to believe that any theories are correct or make correct predictions.

A final negative point was that we witness the graveyard of false theories - a catalogue of non-existent entities and processes.

Positive points in support of scientific realism

Peter Lipton

17th February 2000 : Dr. Martin Kusch


[1] Thesis: Science does not tell us the truth about the world. This is because the notion of 'the truth' as something fixed and unchanging is based on a mistaken view of language.

[2] Two preliminary comments: [a] To give a negative answer to the question 'does science tell the truth?' is not to be antiscientific. [b] Counterintuitive ideas are not necessarily wrong.

[3] My negative answer to the title question is based on a theory about language and truth. Here I can only explain it in a simplified form. My central simplification is to model two different views of the relationship between language and the world on two simple game-scenarios. Call these games Peka and Sepo.

[4] Imagine the following game, called 'Peka.' In order to play the game one needs a big room filled with various objects, two players (A and B) and a stamp (along with an ink pad). Player A leaves the room. While A is out, B takes the stamp, walks around the room and stamps various objects in the room. Some of the stamp-patterns will be openly visible, some will be on object surfaces that are blocked or covered. All of the stamp-patterns will be identical in shape and colour. After a fixed time, A comes back in. His task is to identify all 'pekas.' A 'peka' is an object that has a stamp-pattern.

[5] Now imagine a different game, called 'Sepo.' It needs the same objects-filled room. But it involves three players (A, B, C). No stamp or ink pad are required. While A is out of the room, B and C agree amongst themselves to call three objects 'sepo.' They must choose three objects that they regard as similar. Call these objects, ‘object 1,' 'object 2,' and 'object 3' and their collection the 'sepo-array.' Then A is allowed to return to the room. A is told which three objects are called 'sepo.' Subsequently, A must go around the room, suggest further 'sepos,' and justify his selection to the other players. That is to say, he must find further objects that in his opinion are arguably similar to one or more of the three objects originally picked by B and C. Whether A's choice is correct or not depends on the judgements (i.e. votes) of A, B, and C. When A's suggestion concerning a new sepo is accepted by majority rule (A, B, and C vote), the new object is entered as object 4 into the sepo-array. At the same time, object 1 is dropped from the array. The game then continues relative to the new, changed sepo-array. After a given number of rounds, B (and later C) continues in A's role, but no new array is chosen.

[6] Some general observations:

[6.1] Observations concerning 'Peka.'

[a] Pekas are identified on the basis of the perception of a property. [b] For any given object in the room, it either is stamped or not. And it makes sense to talk about the set of all objects that fall under the concept ‘peka.' 'Peka' has, as philosophers say, a 'fixed extension.'

[c] In identifying pekas, A is making judgements of identity.

[d] Since 'peka' has a fixed extension, it makes sense to speak of player A as making progress. He gets ever closer to identifying all of the pekas.

[6.2] Observations concerning 'Sepo.'

[a] Sepos are determined on the basis of perception and negotiation. A could not correctly identify sepos on his own.

[b] It would be incorrect to say that for any given object in the room, it either is, or is not, a sepo. That is, it would be incorrect to say that for any given object in the room it either is, or is not, similar to the initially chosen three objects. And thus 'sepo' does not have a fixed extension. Why?

[b.1] At any given time, no more than three objects are definitely 'sepo.' And these three objects change with each newly identified sepo.

[b.2] No object is 'sepo' prior to the vote taken by A, B, and C. We cannot say 'many objects in the room are really sepo, the players just haven't found them yet.' The problem with this proposal is that is overlooks the essential openness of similarity judgements. I am not saying that anything can - under normal conditions - be judged to be similar to just anything else. It would take very unusual circumstances indeed for us to judge a bulldozer and a cat to be similar. But such extreme cases aside, we can always go in more than one way. And what that way is, usually depends on our current activities and their goals.

[c] The judgements of A in identifying sepos are judgements concerning similarity, not identity.

[d] Since there is no set properly called 'the set of all sepos,' we cannot meaningfully speak of player A as making progress towards identifying all of the sepos. Remember here that at any given time the sepo-array only has three members. In other words, the array drifts over time. But it does not drift towards finding the final set of all sepos.

[7] Let me now bring in 'truth.' That can be done in two ways.

[7.1] What makes statements 'object n is a peka' or 'object n is a sepo' true?

[a] 'Object n is a peka' is true if and only if n bears the mark of the stamp. This truth is recognition-independent; it holds whether or not we or A recognise it. The truth is determined and fixed. There is a specific number of statements of the form 'object n is a peka' that are true. And that number of statements does not change. It makes sense to say that a successful A discovers more and more truths.

[b] 'Object n is a sepo' is true if and only if n is judged (by the community) to be similar to the three current array elements. The communal similarity judgement is sensitive to empirical properties, but the truth of 'object n is a sepo' is not recognition-independent; it only holds if the players recognise it. Truth is determined and fixed only at a time; but one and the same object can be judged to be a sepo at one time, and later, when the array has changed, not to be a sepo. There is no fixed number of statements with the form 'object n is a sepo' that is true over time. A successful A cannot be said to discover more and more truths. And talk of getting ever closer to the truth does not make sense.

[7.2] What happens to truth if 'peka' or 'sepo’ turned out to be fancy stand-ins for the word 'true?' To make sense of this scenario, we have to slightly change our games. Imagine that the two games were not played with respect to normal physical objects but with respect to sentences. Imagine these written on pieces of paper and distributed over the room. In game Peka, B stamps some of the pieces of paper; in game Sepo, B and C select three sentences as ‘sepo.'

[a] In Peka it makes sense to say that truth is recognition-independent; that 'true' has a fixed extension; and that a successful A gets closer to identifying the truth. A successful A 'tells the truth.'

[b] In Sepo truth is not recognition independent; 'true' does not have a fixed extension, only a statement that is similar to the array-statements is true; and a successful A does not get closer to identifying the truth. A successful A does not 'tell the truth.' There simply is no truth there waiting to be told !

[8] In order to bring the games in contact with life as we know it, we need to make changes in their rules. Here are the most important ones.

[a] In both games, allow for many classificatory words rather than just one.

[b] In both games, allow that the choice of classificatory terms is arbitrary. instead of 'peka' or 'sepo' we could have had any other sound or sequence of letters.

[c] In Peka:

[c.1] Replace the player who stamps with God, nature, or evolution. They sort things into kinds.

[c.2] And replace A with groups of individuals or scientists.

[d] In Sepo:

[d.1] I Drop the stipulation according to which the array can never have more than three elements.

[d.2] Allow that the elements of the array are ‘exemplary cases,' 'paradigms,' or ‘prototypes’, rather than just any classified elements.

[d.3] Change the rules for dropping elements out of the array: which elements are dropped is itself something that must be negotiated amongst the players.

[d.4] Recognise that the exemplars in the array for 'true' are very diverse: we speak of truth in art, truth in painting, true friends, true gold, and true statements in ethics, physics, politics, history, and mathematics. And the same goes for alleged synonyms for 'true' like 'correspondence,' or 'accurate representation.' - What this observation provides is a supplementary way for arguing that there is no truth to be told.

[9] As you will already have suspected, I think that Sepo provides the correct model for thinking about empirical statements, classifications, theories and truth. It is right because it fits with what we know about language learning, with experimental-psychological data concerning similarity judgements, with social-psychological data relating to certainty of decision, with the findings of the history of science, and with how we use language and relate to our environments.

[10] If the Sepo-theory of language and classification I have sketched here is right then: [a] The idea of a sum or set of all truths is incoherent.

[b] Truth cannot simply be defined as ‘correspondence with the facts.' Coherence, consensus and considerations of expediency are also essential to truth.

[c] Truths are the local and passing products of negotiations amongst people who are directed both at their non-social surroundings and at their peers.

[d] There is no truth out there waiting to be told.

[e] It does not follow that we should stop using the words 'true' or 'truth.' What follows is that we should stop using them for purposes for which they are ill-equipped - defending science, for example.

Martin Kusch

Response from Ernie Davidson : Does Science tell the truth? Yes n' No. I may be coming here at a tangent but, I hope the readers get my point. When science gathers a body of knowledge the results are true according to the data and research results at hand. Then later a new piece of data is discovered which when added to the equation throws out the answer.

Let us take Cancer Research. It is virtually drug orientated. Apart from Chemotherapy, which in killing the cancer and sometimes the patient. This means that funding and publicity does not go to other forms of research and treatment. The researchers are dedicated and throughout life, school and their studies and from other professors are moulded into the drug-based research mode. I believe that even in our own democracy we can call this brainwashing.

Some years ago, I read an interesting 2-volume book written by a doctor and a journalist. It stated a conflict of interest in medical research between the companies who spend money researching and the interest in the patient getting better. It gave some specific details of the political and financial interests in an immediate cure not being found. However, from memory it did states that those involved in research were dedicated to finding solutions. Even if those scientists did come across the cancer research, many would naturally have fixed opinions and probably even refuse to accept the research had any credibility.

In the book, it stated that key drug companies financed Cancer research, since they were developing treatments. However their research excluded and even suppressed other forms of research that had a track record of research. The research claims that adding vitamin B17 to the diet could prevent and treat Cancer. This has dropped out of a great deal of our modern diets.

Here's the web-site :

Let's take medicine in general: here are some interesting case histories.

Migraine : The patient had chronic headaches that must have been Migraine. These would last for several days. The Chinese doctor examined him for a few seconds and told him to lower his air-conditioner and gave him some herbs for 3 days. The migraine never returned.

High Blood Pressure : I saw a person's blood pressure back to normal where it still remained years later, in just three weeks of Chinese Medical treatment. Apparently, it contained a few animals. The Chinese Doctor will usually measure both arms as the blood pressure differs in these.

Yuppie 'Flu : There are several causes but the symptoms can be treated very simply using Traditional Chinese Medicine, with very little change to the routine, but the doctor may advise a reduction of alcohol.

We are taught that our own doctors know best, even when they can give a qualified don't know in how to treat the patient. Of course there are treatments that we wouldn't like too much, such as :-

Dog penis : for kidney complaints, and part of potency treatment (they use other types). I'm told this is rather salty.

Snake : for potency and treatment of rheumatism (with excellent results). This tastes like chicken. One Chinese chef says that therefore next time you eat chicken, remember it tastes like snake.

Lizard : for asthma

Sheep placenta : for beautifying a woman's skin. (I've seen it work).

Donkey vagina : For treating women who are ill after having a baby. Chinese doctors often call it butterfly due to its shape.

Donkey skin juice : For anaemia I spoke to someone who told me it worked for her.

Teas : Different teas for different seasons. Many old people swear by winter teas in that they are warm enough to wear less clothes in the harsh winters.

Because of our culture and backgrounds, our medical research would not look into these items as a first point of research. I may seem to be going off the point, but this is to illustrate how we sometimes treat information when analysing facts.

Here's a hypothetical case. A photograph of a politician talking to a known prostitute is printed in the press. Basically it is nothing but a photograph of a politician who appears to be talking to a known prostitute. Many readers will then start to imagine what he was doing, although the only information they have is the photo. The newspaper will probably carefully comment on the photo to mislead the readers. 'Politician caught on camera with prostitute". However, the politician was out with his wife who is not in the picture as she was in one of the shops, and the prostitute recognised him and went to complain about the fact she had been on a council waiting list for 2 years……

Err.. that's what the politician said. The photographer by chance spotted the conversation, clicked the camera and then ran. He was true in that he saw a politician talking to a known prostitute.

Any comments ?

Ernie Davidson

Bob Williams : The first thing that struck me about both papers was that there was that there was not a clear attempt to define the very topic. If we ask "does science tell the truth," we have an obligation to make it clear as to what this is supposed to mean. Does it mean that all sciences must be 100% correct in the predictions made by all practising scientists in each discipline? I don't think many of us would buy into that. Does it mean that "science" is correct more than half of the time? 90% of the time? 99.92% of the time? These guys didn't seem to want to say, but were more interested in abstractions and contortions.

They did not make it at all clear what the term "science" means? Physics? Psychology? Archeology? Geography? Geology? Astronomy? Medicine? There are big differences as to how quantitative these are and about how they must deal with their subject matter.

Truth is the subject that philosophers love to discuss, but always in the abstract and, in my observation, usually to no useful end point. If a physicist says F = ma, is he making a statement of truth? If he puts values into the equation, is the statement more or less true? I accept that we know what this equation says and how to use it and how to understand and deal with any associated errors.

If an astronomer says the age of the universe is 12 billion years and another says 14 billion, they are doing so from the perspective of evidence that they believe is "best." We know that both are guessing, but doing so on the basis of available information and knowledge. Does this bother a philosopher? Does he say that they are not being "true?"

Much of science is based on statistics and can only be understood in terms of statistics. Within these areas, there are large exceptions to the general trends. If we look at growth rates, or disease resistance, or nerve conduction velocities, we can make very useful statistical observations, but there are data all over the place and each datum may be completely correct. So, in this context, what is "truth?"

" ... there is no sense in which his theories are true or false; they are just empirically adequate (or not)." : This simply defines away the arguments of absolute truth, which are probably meaningless in real world applications.

" ... the under-determination of theory by data." : It is not clear to me what is being argued. Is "truth" taken to mean only the accurate and infallible statement of scientific principles in the form of theories? For example, is it a scientific truth when a researcher identifies a specific virus as the cause of a disease? When a scientist measures the wavelength of light from a laser source? When DNA is examined and linked to another DNA source? When moon rocks are chemically analyzed and reported?

Look at these per the "alternate theory" concept. Do we have alternate theories about the causes of colds and AIDS? I thought there was pretty good agreement as to which viruses caused the diseases. Do we have alternate theories about laser light wavelengths? Does one scientist say that the light is red and the other says it is blue? Do we have people arguing that DNA links are coincidental, or have no meaning, or are falsely observed, or in some other way dispute them? Do we have people telling us that the moon rocks that have been studied are really cheese, as was long expected?

It seems to me that Martin Kusch wants to discuss language and not science. But, if he wants to discuss language, why doesn't he just proceed to define his terms clearly?

I started reading the "game" story and could not force myself to read completely through it. This kind of discussion strikes me as contrived for no useful purpose, by a guy who does not address the basic issues of what he means by science and truth.

If one removes the abstractions and looks at the history of scientific thought in various disciplines, he must conclude that there has been an evolution of increasing understanding of the subjects and that the theories used by scientists have evolved to more useful and defendable forms. It is the nature of science that we look at some matters that are so difficult to study that we may never have strong proofs of what we are inclined to believe, but we seem to continually build stronger cases for our understandings of how all of science relates (one part to another). Meanwhile, the ultimate test of how well we have done is to see if we have been able to use our knowledge productively to build machines, fight disease, improve health, and to control our environments. Whatever we accept as truth has worked well enough to allow us to build complex machines, to communicate over great distances, and to do the thousands of other things that were not possible before we reached our present level of scientific understanding of nature.

Bob Williams

Previous Article in Current Issue (Commensal 103)
Index to Current Issue (Commensal 103)